[AMDGPU] Make amdgcn.s.buffer.load a memory-reading intrinsic

[krzysz00]

Summary

Change the s.buffer.load to have a memory read effect, correctly
representing that it reads memory and matching the backend's
representation of the intrinsic.

This is forked out of D146761, a larger refactoring of the buffer
intrinsics.

On its own, this commit may have performance implications for users of
s.buffer.load (who are, to my knowledge, mainly Mesa) who may be
relying on that lack of memory effect to hoist loads from constant,
invariant buffers out of loops, for example.

However, when the full buffer refactoring lands, s.buffer.load's new
form will take a pointer-valued argument and thus, the constant-ness
of the buffer will be expressible using annotations as the callsite,
such as "noalias", so this fixup should not have a long-term
performance impact.

[mareko]

Mesa sets "invariant.load" metadata on all "s.buffer.load" call sites. There are no pointers, only descriptors. Is that sufficient to get the readnone behavior?

[krzysz00]

The plan here is to introduce new intrinsics that do use pointers - that's a future patch in the series - so that we can properly reason about things like buffer alias analysis. So this is an intermediate step that will break perf until there's migration to the new intrinsics.

[mareko]

We can't break perf, not even temporarily.

[foad]

I had a quick look at some codegen differences this caused in Vulkan shaders compiled by LLPC. The biggest difference was that the DAG nodes for scalar loads now have a chain operand, which has a big effect on DAG scheduling.

[arsenm]

In D147245#4234726, @mareko wrote:

Mesa sets "invariant.load" metadata on all "s.buffer.load" call sites. There are no pointers, only descriptors. Is that sufficient to get the readnone behavior?

I don't know if that's even valid. Without looking I'm 95% sure that won't do anything. Ideally that would work

[arsenm]

In D147245#4236143, @foad wrote:

I had a quick look at some codegen differences this caused in Vulkan shaders compiled by LLPC. The biggest difference was that the DAG nodes for scalar loads now have a chain operand, which has a big effect on DAG scheduling.

Oh right, I forgot about the DAG chaining. The chain must be determined by the declaration (otherwise code would have to randomly support chained and unchained variants of memory intrinsics)

[krzysz00]

Ok, so, just to clarify:

1. There's an undocumented invariant that the buffer you pass in to s.buffer.load is constant for the duration of the program.
2. A bunch of folks are relying on that invariant for performance.
3. That aside, s.buffer.load is, more or less, raw.buffer.load

Given that those things are true, I'm going to back out of this change to s.buffer.load, and we'll leave all these codepaths untouched during the buffer intrinsic updates for now. Once address space 7 properly exists, folks'll be able to migrate to that system and use a typical load with the appropriate "this points to constant memory" annotations, so we may end up not needing this special intrinsic.

So, I'll go ahead and file a new revision to update the docs.

[krzysz00]

From what I can tell, if !invariant.load is supported for load-like intrinsics, that's not documented anywhere. It might be neat to make that work, but we can cross that bridge when we get there.

On the other hand, I'm imaging the following sequence of rewrites:

%v = load T, ptr addrspace(7) %const.buffer, !invariant.load

=> (late in the frontend)

%buf, %offset = ...(%const.buffer)
%v = call T @llvm.amdgcn.raw.buffer.load(%buf, i32 %off. [...]) !invariant.load

=> (in IR translation)

INTRINSIC_W_SIDE_EFFECT(@llvm.amdgcn.raw.buffer.load, ...) :: (dereferencable invariant load from BufferAddr(%buf, %off) ....)

=>(because the load was marked invariant, assuming %off is uniform)

AMDGCN_S_BUFFER_LOAD %buf, %off, ... :: (...)

[foad]

From what I can tell, if !invariant.load is supported for load-like intrinsics, that's not documented anywhere. It might be neat to make that work, but we can cross that bridge when we get there.

As far as I know it's only supported ad hoc for some AMDGPU load-like intrinsics: D119739

[krzysz00]

This here isn't going to work, closing to unclutter everyone's review lists

llvm/include/llvm/IR/IntrinsicsAMDGPU.td

 
 // Generate a buffer_load instruction that may be optimized to s_buffer_load if
 // the offset argument is uniform.
 def int_amdgcn_s_buffer_load : DefaultAttrsIntrinsic <
   [llvm_any_ty],
   [llvm_v4i32_ty,     // rsrc(SGPR)
    llvm_i32_ty,       // byte offset
    llvm_i32_ty],      // cachepolicy(imm; bit 0 = glc, bit 2 = dlc)
-  [IntrNoMem, ImmArg<ArgIndex<2>>]>,
+  [IntrReadMem, ImmArg<ArgIndex<2>>]>,
   AMDGPURsrcIntrinsic<0>;
 
 class AMDGPUBufferStore<LLVMType data_ty = llvm_any_ty> : DefaultAttrsIntrinsic <
   [],
   [data_ty,          // vdata(VGPR)
    llvm_v4i32_ty,     // rsrc(SGPR)
    llvm_i32_ty,       // vindex(VGPR)
    llvm_i32_ty,       // offset(SGPR/VGPR/imm)

llvm/lib/Target/AMDGPU/AMDGPULegalizerInfo.cpp

 #include "AMDGPUTargetMachine.h"
 #include "SIMachineFunctionInfo.h"
 #include "Utils/AMDGPUBaseInfo.h"
 #include "llvm/ADT/ScopeExit.h"
 #include "llvm/BinaryFormat/ELF.h"
 #include "llvm/CodeGen/GlobalISel/LegalizerHelper.h"
 #include "llvm/CodeGen/GlobalISel/MIPatternMatch.h"
 #include "llvm/CodeGen/GlobalISel/MachineIRBuilder.h"
+#include "llvm/CodeGen/MachineMemOperand.h"
 #include "llvm/IR/DiagnosticInfo.h"
 #include "llvm/IR/IntrinsicsAMDGPU.h"
 #include "llvm/IR/IntrinsicsR600.h"
 
 #define DEBUG_TYPE "amdgpu-legalinfo"
 
 using namespace llvm;
 using namespace LegalizeActions;
bool AMDGPULegalizerInfo::legalizeImageIntrinsic(
   B.buildConcatVectors(DstReg, ResultRegs);
   return true;
 }
 
 bool AMDGPULegalizerInfo::legalizeSBufferLoad(
   LegalizerHelper &Helper, MachineInstr &MI) const {
   MachineIRBuilder &B = Helper.MIRBuilder;
   GISelChangeObserver &Observer = Helper.Observer;
+  MachineFunction &MF = B.getMF();
 
   Register Dst = MI.getOperand(0).getReg();
   LLT Ty = B.getMRI()->getType(Dst);
   unsigned Size = Ty.getSizeInBits();
-  MachineFunction &MF = B.getMF();
 
   Observer.changingInstr(MI);
 
   if (shouldBitcastLoadStoreType(ST, Ty, LLT::scalar(Size))) {
     Ty = getBitcastRegisterType(Ty);
     Helper.bitcastDst(MI, Ty, 0);
     Dst = MI.getOperand(0).getReg();
     B.setInsertPt(B.getMBB(), MI);
   }
 
   // FIXME: We don't really need this intermediate instruction. The intrinsic
   // should be fixed to have a memory operand. Since it's readnone, we're not
   // allowed to add one.
   MI.setDesc(B.getTII().get(AMDGPU::G_AMDGPU_S_BUFFER_LOAD));
   MI.removeOperand(1); // Remove intrinsic ID
 
-  // FIXME: When intrinsic definition is fixed, this should have an MMO already.
-  // TODO: Should this use datalayout alignment?
-  const unsigned MemSize = (Size + 7) / 8;
-  const Align MemAlign(4);
-  MachineMemOperand *MMO = MF.getMachineMemOperand(
-      MachinePointerInfo(),
-      MachineMemOperand::MOLoad | MachineMemOperand::MODereferenceable |
-          MachineMemOperand::MOInvariant,
-      MemSize, MemAlign);
-  MI.addMemOperand(MF, MMO);
+  MachineMemOperand *MMO = *MI.memoperands_begin();
+  MMO->setFlags(MMO->getFlags() | MachineMemOperand::MOInvariant);
+
+  // TODO: Consider using the alignment of the underlying data type here.
+  const Align AssertedAlign(4);
+  MachineMemOperand *NewAlignMMO = MF.getMachineMemOperand(
+      MMO->getPointerInfo(), MMO->getFlags(), MMO->getSize(), AssertedAlign);
+  MMO->refineAlignment(NewAlignMMO);
 
   // There are no 96-bit result scalar loads, but widening to 128-bit should
   // always be legal. We may need to restore this to a 96-bit result if it turns
   // out this needs to be converted to a vector load during RegBankSelect.
   if (!isPowerOf2_32(Size)) {
     if (Ty.isVector())
       Helper.moreElementsVectorDst(MI, getPow2VectorType(Ty), 0);
     else

llvm/lib/Target/AMDGPU/AMDGPURegisterBankInfo.cpp

bool AMDGPURegisterBankInfo::applyMappingSBufferLoad(
 
   unsigned MMOOffset = setBufferOffsets(B, MI.getOperand(2).getReg(), VOffset,
                                         SOffset, ImmOffset, Alignment);
 
   // TODO: 96-bit loads were widened to 128-bit results. Shrink the result if we
   // can, but we need to track an MMO for that.
   const unsigned MemSize = (Ty.getSizeInBits() + 7) / 8;
   const Align MemAlign(4); // FIXME: ABI type alignment?
+  MachineMemOperand *OrigMMO = *MI.memoperands_begin();
   MachineMemOperand *BaseMMO = MF.getMachineMemOperand(
-    MachinePointerInfo(),
+      OrigMMO->getPointerInfo(),
       MachineMemOperand::MOLoad | MachineMemOperand::MODereferenceable |
           MachineMemOperand::MOInvariant,
       MemSize, MemAlign);
   if (MMOOffset != 0)
     BaseMMO = MF.getMachineMemOperand(BaseMMO, MMOOffset, MemSize);
 
   // If only the offset is divergent, emit a MUBUF buffer load instead. We can
   // assume that the buffer is unswizzled.
 
   Register RSrc = MI.getOperand(1).getReg();
   Register VIndex = B.buildConstant(S32, 0).getReg(0);

llvm/lib/Target/AMDGPU/SIISelLowering.h

SDValue getPreloadedValue(SelectionDAG &DAG,
                             AMDGPUFunctionArgInfo::PreloadedValue) const;
 
   SDValue LowerGlobalAddress(AMDGPUMachineFunction *MFI, SDValue Op,
                              SelectionDAG &DAG) const override;
   SDValue lowerImplicitZextParam(SelectionDAG &DAG, SDValue Op,
                                  MVT VT, unsigned Offset) const;
   SDValue lowerImage(SDValue Op, const AMDGPU::ImageDimIntrinsicInfo *Intr,
                      SelectionDAG &DAG, bool WithChain) const;
-  SDValue lowerSBuffer(EVT VT, SDLoc DL, SDValue Rsrc, SDValue Offset,
-                       SDValue CachePolicy, SelectionDAG &DAG) const;
+  SDValue lowerSBuffer(EVT VT, SDLoc DL, SDValue Chain, SDValue Rsrc,
+                       SDValue Offset, SDValue CachePolicy,
+                       SelectionDAG &DAG) const;
 
   SDValue lowerRawBufferAtomicIntrin(SDValue Op, SelectionDAG &DAG,
                                      unsigned NewOpcode) const;
   SDValue lowerStructBufferAtomicIntrin(SDValue Op, SelectionDAG &DAG,
                                         unsigned NewOpcode) const;
 
   SDValue lowerWorkitemID(SelectionDAG &DAG, SDValue Op, unsigned Dim,
                           const ArgDescriptor &ArgDesc) const;

llvm/lib/Target/AMDGPU/SIISelLowering.cpp

if (Size > 32) {
       return NumIntermediates;
     }
   }
 
   return TargetLowering::getVectorTypeBreakdownForCallingConv(
     Context, CC, VT, IntermediateVT, NumIntermediates, RegisterVT);
 }
 
-static EVT memVTFromLoadIntrData(Type *Ty, unsigned MaxNumLanes) {
+static EVT memVTFromLoadIntrData(Type *Ty, const DataLayout &DL,
+                                 unsigned MaxNumLanes) {
   assert(MaxNumLanes != 0);
+  if (auto *PT = dyn_cast<PointerType>(Ty))
+    return EVT::getIntegerVT(Ty->getContext(),
+                             DL.getPointerSizeInBits(PT->getAddressSpace()));
 
   if (auto *VT = dyn_cast<FixedVectorType>(Ty)) {
     unsigned NumElts = std::min(MaxNumLanes, VT->getNumElements());
-    return EVT::getVectorVT(Ty->getContext(),
-                            EVT::getEVT(VT->getElementType()),
-                            NumElts);
+    Type *ElemTy = VT->getElementType();
+    EVT ElemTyVT = EVT::getEVT(ElemTy);
+    if (auto *PT = dyn_cast<PointerType>(ElemTy))
+      ElemTyVT = EVT::getIntegerVT(
+          Ty->getContext(), DL.getPointerSizeInBits(PT->getAddressSpace()));
+    return EVT::getVectorVT(Ty->getContext(), ElemTyVT, NumElts);
   }
 
   return EVT::getEVT(Ty);
 }
 
 // Peek through TFE struct returns to only use the data size.
-static EVT memVTFromLoadIntrReturn(Type *Ty, unsigned MaxNumLanes) {
+static EVT memVTFromLoadIntrReturn(Type *Ty, const DataLayout &DL,
+                                   unsigned MaxNumLanes) {
   auto *ST = dyn_cast<StructType>(Ty);
   if (!ST)
-    return memVTFromLoadIntrData(Ty, MaxNumLanes);
+    return memVTFromLoadIntrData(Ty, DL, MaxNumLanes);
 
   // TFE intrinsics return an aggregate type.
   assert(ST->getNumContainedTypes() == 2 &&
          ST->getContainedType(1)->isIntegerTy(32));
-  return memVTFromLoadIntrData(ST->getContainedType(0), MaxNumLanes);
+  return memVTFromLoadIntrData(ST->getContainedType(0), DL, MaxNumLanes);
 }
 
 bool SITargetLowering::getTgtMemIntrinsic(IntrinsicInfo &Info,
                                           const CallInst &CI,
                                           MachineFunction &MF,
                                           unsigned IntrID) const {
+  const DataLayout &DL = MF.getDataLayout();
   Info.flags = MachineMemOperand::MONone;
   if (CI.hasMetadata(LLVMContext::MD_invariant_load))
     Info.flags |= MachineMemOperand::MOInvariant;
 
   if (const AMDGPU::RsrcIntrinsic *RsrcIntr =
           AMDGPU::lookupRsrcIntrinsic(IntrID)) {
     AttributeList Attr = Intrinsic::getAttributes(CI.getContext(),
                                                   (Intrinsic::ID)IntrID);
     MemoryEffects ME = Attr.getMemoryEffects();
     if (ME.doesNotAccessMemory())
       return false;
 
     // TODO: Should images get their own address space?
     Info.fallbackAddressSpace = AMDGPUAS::BUFFER_FAT_POINTER;
 
     if (RsrcIntr->IsImage)
       Info.align.reset();
 
     Info.flags |= MachineMemOperand::MODereferenceable;
     if (ME.onlyReadsMemory()) {
-      unsigned MaxNumLanes = 4;
+      unsigned MaxNumLanes = IntrID == Intrinsic::amdgcn_s_buffer_load ? 16 : 4;
 
       if (RsrcIntr->IsImage) {
         const AMDGPU::ImageDimIntrinsicInfo *Intr
           = AMDGPU::getImageDimIntrinsicInfo(IntrID);
         const AMDGPU::MIMGBaseOpcodeInfo *BaseOpcode =
           AMDGPU::getMIMGBaseOpcodeInfo(Intr->BaseOpcode);
 
         if (!BaseOpcode->Gather4) {
           // If this isn't a gather, we may have excess loaded elements in the
           // IR type. Check the dmask for the real number of elements loaded.
           unsigned DMask
             = cast<ConstantInt>(CI.getArgOperand(0))->getZExtValue();
           MaxNumLanes = DMask == 0 ? 1 : llvm::popcount(DMask);
         }
       }
 
-      Info.memVT = memVTFromLoadIntrReturn(CI.getType(), MaxNumLanes);
+      Info.memVT = memVTFromLoadIntrReturn(CI.getType(), DL, MaxNumLanes);
 
       // FIXME: What does alignment mean for an image?
       Info.opc = ISD::INTRINSIC_W_CHAIN;
       Info.flags |= MachineMemOperand::MOLoad;
     } else if (ME.onlyWritesMemory()) {
       Info.opc = ISD::INTRINSIC_VOID;
 
       Type *DataTy = CI.getArgOperand(0)->getType();
       if (RsrcIntr->IsImage) {
         unsigned DMask = cast<ConstantInt>(CI.getArgOperand(1))->getZExtValue();
         unsigned DMaskLanes = DMask == 0 ? 1 : llvm::popcount(DMask);
-        Info.memVT = memVTFromLoadIntrData(DataTy, DMaskLanes);
+        Info.memVT = memVTFromLoadIntrData(DataTy, DL, DMaskLanes);
       } else
         Info.memVT = EVT::getEVT(DataTy);
 
       Info.flags |= MachineMemOperand::MOStore;
     } else {
       // Atomic
       Info.opc = CI.getType()->isVoidTy() ? ISD::INTRINSIC_VOID :
                                             ISD::INTRINSIC_W_CHAIN;
SDValue SITargetLowering::lowerImage(SDValue Op,
   if (BaseOpcode->Store)
     return SDValue(NewNode, 0);
   return constructRetValue(DAG, NewNode,
                            OrigResultTypes, IsTexFail,
                            Subtarget->hasUnpackedD16VMem(), IsD16,
                            DMaskLanes, NumVDataDwords, DL);
 }
 
-SDValue SITargetLowering::lowerSBuffer(EVT VT, SDLoc DL, SDValue Rsrc,
-                                       SDValue Offset, SDValue CachePolicy,
+SDValue SITargetLowering::lowerSBuffer(EVT VT, SDLoc DL, SDValue Chain,
+                                       SDValue Rsrc, SDValue Offset,
+                                       SDValue CachePolicy,
                                        SelectionDAG &DAG) const {
   MachineFunction &MF = DAG.getMachineFunction();
 
   const DataLayout &DataLayout = DAG.getDataLayout();
   Align Alignment =
       DataLayout.getABITypeAlign(VT.getTypeForEVT(*DAG.getContext()));
 
   MachineMemOperand *MMO = MF.getMachineMemOperand(
       MachinePointerInfo(),
       MachineMemOperand::MOLoad | MachineMemOperand::MODereferenceable |
           MachineMemOperand::MOInvariant,
       VT.getStoreSize(), Alignment);
 
   if (!Offset->isDivergent()) {
-    SDValue Ops[] = {
-        Rsrc,
+    SDValue Ops[] = {Chain, Rsrc,
                      Offset, // Offset
-        CachePolicy
-    };
+                     CachePolicy};
 
     // Widen vec3 load to vec4.
     if (VT.isVector() && VT.getVectorNumElements() == 3) {
       EVT WidenedVT =
           EVT::getVectorVT(*DAG.getContext(), VT.getVectorElementType(), 4);
       auto WidenedOp = DAG.getMemIntrinsicNode(
-          AMDGPUISD::SBUFFER_LOAD, DL, DAG.getVTList(WidenedVT), Ops, WidenedVT,
+          AMDGPUISD::SBUFFER_LOAD, DL, DAG.getVTList(WidenedVT, MVT::Other),
+          Ops, WidenedVT,
           MF.getMachineMemOperand(MMO, 0, WidenedVT.getStoreSize()));
       auto Subvector = DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, VT, WidenedOp,
                                    DAG.getVectorIdxConstant(0, DL));
-      return Subvector;
+      auto Merged =
+          DAG.getMergeValues({Subvector, SDValue(WidenedOp.getNode(), 1)}, DL);
+      return Merged;
     }
 
     return DAG.getMemIntrinsicNode(AMDGPUISD::SBUFFER_LOAD, DL,
-                                   DAG.getVTList(VT), Ops, VT, MMO);
+                                   DAG.getVTList(VT, MVT::Other), Ops, VT, MMO);
   }
 
   // We have a divergent offset. Emit a MUBUF buffer load instead. We can
   // assume that the buffer is unswizzled.
   SmallVector<SDValue, 4> Loads;
   unsigned NumLoads = 1;
   MVT LoadVT = VT.getSimpleVT();
   unsigned NumElts = LoadVT.isVector() ? LoadVT.getVectorNumElements() : 1;
   assert((LoadVT.getScalarType() == MVT::i32 ||
           LoadVT.getScalarType() == MVT::f32));
 
   if (NumElts == 8 || NumElts == 16) {
     NumLoads = NumElts / 4;
     LoadVT = MVT::getVectorVT(LoadVT.getScalarType(), 4);
   }
 
-  SDVTList VTList = DAG.getVTList({LoadVT, MVT::Glue});
+  SDVTList VTList = DAG.getVTList({LoadVT, MVT::Other});
   SDValue Ops[] = {
-      DAG.getEntryNode(),                               // Chain
-      Rsrc,                                             // rsrc
-      DAG.getConstant(0, DL, MVT::i32),                 // vindex
-      {},                                               // voffset
-      {},                                               // soffset
-      {},                                               // offset
-      CachePolicy,                                      // cachepolicy
-      DAG.getTargetConstant(0, DL, MVT::i1),            // idxen
+      Chain,
+      Rsrc,                                  // rsrc
+      DAG.getConstant(0, DL, MVT::i32),      // vindex
+      {},                                    // voffset
+      {},                                    // soffset
+      {},                                    // offset
+      CachePolicy,                           // cachepolicy
+      DAG.getTargetConstant(0, DL, MVT::i1), // idxen
   };
 
   // Use the alignment to ensure that the required offsets will fit into the
   // immediate offsets.
   setBufferOffsets(Offset, DAG, &Ops[3],
                    NumLoads > 1 ? Align(16 * NumLoads) : Align(4));
 
   uint64_t InstOffset = cast<ConstantSDNode>(Ops[5])->getZExtValue();
+
   for (unsigned i = 0; i < NumLoads; ++i) {
     Ops[5] = DAG.getTargetConstant(InstOffset + 16 * i, DL, MVT::i32);
     Loads.push_back(getMemIntrinsicNode(AMDGPUISD::BUFFER_LOAD, DL, VTList, Ops,
                                         LoadVT, MMO, DAG));
   }
 
-  if (NumElts == 8 || NumElts == 16)
-    return DAG.getNode(ISD::CONCAT_VECTORS, DL, VT, Loads);
+  if (NumElts == 8 || NumElts == 16) {
+    SDValue Result = DAG.getNode(ISD::CONCAT_VECTORS, DL, VT, Loads);
+    SmallVector<SDValue> Chains;
+    for (SDValue Load : Loads)
+      Chains.emplace_back(Load.getNode(), 1);
+    SDValue NewChain = DAG.getTokenFactor(DL, Chains);
+    return DAG.getMergeValues({Result, NewChain}, DL);
+  }
 
   return Loads[0];
 }
 
 SDValue SITargetLowering::lowerWorkitemID(SelectionDAG &DAG, SDValue Op,
                                           unsigned Dim,
                                           const ArgDescriptor &Arg) const {
   SDLoc SL(Op);
case Intrinsic::amdgcn_workitem_id_x:
     return lowerWorkitemID(DAG, Op, 0, MFI->getArgInfo().WorkItemIDX);
   case Intrinsic::amdgcn_workitem_id_y:
     return lowerWorkitemID(DAG, Op, 1, MFI->getArgInfo().WorkItemIDY);
   case Intrinsic::amdgcn_workitem_id_z:
     return lowerWorkitemID(DAG, Op, 2, MFI->getArgInfo().WorkItemIDZ);
   case Intrinsic::amdgcn_wavefrontsize:
     return DAG.getConstant(MF.getSubtarget<GCNSubtarget>().getWavefrontSize(),
                            SDLoc(Op), MVT::i32);
-  case Intrinsic::amdgcn_s_buffer_load: {
-    unsigned CPol = cast<ConstantSDNode>(Op.getOperand(3))->getZExtValue();
-    if (CPol & ~AMDGPU::CPol::ALL)
-      return Op;
-    return lowerSBuffer(VT, DL, Op.getOperand(1), Op.getOperand(2), Op.getOperand(3),
-                        DAG);
-  }
   case Intrinsic::amdgcn_fdiv_fast:
     return lowerFDIV_FAST(Op, DAG);
   case Intrinsic::amdgcn_sin:
     return DAG.getNode(AMDGPUISD::SIN_HW, DL, VT, Op.getOperand(1));
 
   case Intrinsic::amdgcn_cos:
     return DAG.getNode(AMDGPUISD::COS_HW, DL, VT, Op.getOperand(1));
 
SDValue Ops[] = {
       M->getOperand(0), // Chain
       M->getOperand(2), // Ptr
       M->getOperand(3)  // Value
     };
 
     return DAG.getMemIntrinsicNode(Opc, SDLoc(Op), M->getVTList(), Ops,
                                    M->getMemoryVT(), M->getMemOperand());
   }
+  case Intrinsic::amdgcn_s_buffer_load: {
+    unsigned CPol = cast<ConstantSDNode>(Op.getOperand(4))->getZExtValue();
+    if (CPol & ~AMDGPU::CPol::ALL)
+      return Op;
+    SDValue Chain = Op->getOperand(0);
+    EVT VT = Op->getValueType(0);
+    SDValue Rsrc = Op.getOperand(2);
+    return lowerSBuffer(VT, DL, Chain, Rsrc, Op.getOperand(3), Op.getOperand(4),
+                        DAG);
+  }
   case Intrinsic::amdgcn_buffer_load:
   case Intrinsic::amdgcn_buffer_load_format: {
     unsigned Glc = cast<ConstantSDNode>(Op.getOperand(5))->getZExtValue();
     unsigned Slc = cast<ConstantSDNode>(Op.getOperand(6))->getZExtValue();
     unsigned IdxEn = getIdxEn(Op.getOperand(3));
     SDValue Ops[] = {
       Op.getOperand(0), // Chain
       Op.getOperand(2), // rsrc

llvm/lib/Target/AMDGPU/SIInstrInfo.td

 //===----------------------------------------------------------------------===//
 // SI DAG Nodes
 //===----------------------------------------------------------------------===//
 
 def AMDGPUclamp : SDNode<"AMDGPUISD::CLAMP", SDTFPUnaryOp>;
 
 def SIsbuffer_load : SDNode<"AMDGPUISD::SBUFFER_LOAD",
   SDTypeProfile<1, 3, [SDTCisVT<1, v4i32>, SDTCisVT<2, i32>, SDTCisVT<3, i32>]>,
-  [SDNPMayLoad, SDNPMemOperand]
+  [SDNPMayLoad, SDNPMemOperand, SDNPHasChain]
 >;
 
 def SIds_ordered_count : SDNode<"AMDGPUISD::DS_ORDERED_COUNT",
   SDTypeProfile<1, 2, [SDTCisVT<0, i32>, SDTCisVT<1, i32>, SDTCisVT<2, i16>]>,
   [SDNPMayLoad, SDNPMayStore, SDNPMemOperand, SDNPHasChain, SDNPInGlue]
 >;
 
 def SDTAtomic2_f32 : SDTypeProfile<1, 2, [

llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-llvm.amdgcn.s.buffer.load.mir

bb.0:
     ; GCN: liveins: $sgpr0_sgpr1_sgpr2_sgpr3
     ; GCN-NEXT: {{  $}}
     ; GCN-NEXT: [[COPY:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
     ; GCN-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
     ; GCN-NEXT: [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:_(s32) = G_AMDGPU_S_BUFFER_LOAD [[COPY]](<4 x s32>), [[C]](s32), 0 :: (dereferenceable invariant load (s32))
     ; GCN-NEXT: S_ENDPGM 0, implicit [[AMDGPU_S_BUFFER_LOAD]](s32)
     %0:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
     %1:_(s32) = G_CONSTANT i32 0
-    %2:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.s.buffer.load), %0, %1, 0
+    %2:_(s32) = G_INTRINSIC_W_SIDE_EFFECTS intrinsic(@llvm.amdgcn.s.buffer.load), %0, %1, 0 :: (dereferenceable invariant load (s32), align 4)
     S_ENDPGM 0, implicit %2
 
 ...
 
 ---
 name: s_buffer_load_v3s32
 body:             |
   bb.0:
     liveins: $sgpr0_sgpr1_sgpr2_sgpr3
 
     ; GCN-LABEL: name: s_buffer_load_v3s32
     ; GCN: liveins: $sgpr0_sgpr1_sgpr2_sgpr3
     ; GCN-NEXT: {{  $}}
     ; GCN-NEXT: [[COPY:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
     ; GCN-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
     ; GCN-NEXT: [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:_(<4 x s32>) = G_AMDGPU_S_BUFFER_LOAD [[COPY]](<4 x s32>), [[C]](s32), 0 :: (dereferenceable invariant load (s96), align 4)
     ; GCN-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_S_BUFFER_LOAD]](<4 x s32>)
     ; GCN-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[UV]](s32), [[UV1]](s32), [[UV2]](s32)
     ; GCN-NEXT: S_ENDPGM 0, implicit [[BUILD_VECTOR]](<3 x s32>)
     %0:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
     %1:_(s32) = G_CONSTANT i32 0
-    %2:_(<3 x s32>) = G_INTRINSIC intrinsic(@llvm.amdgcn.s.buffer.load), %0, %1, 0
+    %2:_(<3 x s32>) = G_INTRINSIC_W_SIDE_EFFECTS intrinsic(@llvm.amdgcn.s.buffer.load), %0, %1, 0 :: (dereferenceable invariant load (s96), align 1)
     S_ENDPGM 0, implicit %2
 
 ...
 
 ---
 name: s_buffer_load_v3p3
 body:             |
   bb.0:
     liveins: $sgpr0_sgpr1_sgpr2_sgpr3
 
     ; GCN-LABEL: name: s_buffer_load_v3p3
     ; GCN: liveins: $sgpr0_sgpr1_sgpr2_sgpr3
     ; GCN-NEXT: {{  $}}
     ; GCN-NEXT: [[COPY:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
     ; GCN-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
     ; GCN-NEXT: [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:_(<4 x s32>) = G_AMDGPU_S_BUFFER_LOAD [[COPY]](<4 x s32>), [[C]](s32), 0 :: (dereferenceable invariant load (s96), align 4)
     ; GCN-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_S_BUFFER_LOAD]](<4 x s32>)
     ; GCN-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[UV]](s32), [[UV1]](s32), [[UV2]](s32)
     ; GCN-NEXT: [[BITCAST:%[0-9]+]]:_(<3 x p3>) = G_BITCAST [[BUILD_VECTOR]](<3 x s32>)
     ; GCN-NEXT: S_ENDPGM 0, implicit [[BITCAST]](<3 x p3>)
     %0:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
     %1:_(s32) = G_CONSTANT i32 0
-    %2:_(<3 x p3>) = G_INTRINSIC intrinsic(@llvm.amdgcn.s.buffer.load), %0, %1, 0
+    %2:_(<3 x p3>) = G_INTRINSIC_W_SIDE_EFFECTS intrinsic(@llvm.amdgcn.s.buffer.load), %0, %1, 0 :: (dereferenceable invariant load (s96), align 1)
     S_ENDPGM 0, implicit %2
 
 ...
 
 ---
 name: s_buffer_load_v6s16
 body:             |
   bb.0:
     liveins: $sgpr0_sgpr1_sgpr2_sgpr3
 
     ; GCN-LABEL: name: s_buffer_load_v6s16
     ; GCN: liveins: $sgpr0_sgpr1_sgpr2_sgpr3
     ; GCN-NEXT: {{  $}}
     ; GCN-NEXT: [[COPY:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
     ; GCN-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
     ; GCN-NEXT: [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:_(<4 x s32>) = G_AMDGPU_S_BUFFER_LOAD [[COPY]](<4 x s32>), [[C]](s32), 0 :: (dereferenceable invariant load (s96), align 4)
     ; GCN-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_S_BUFFER_LOAD]](<4 x s32>)
     ; GCN-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[UV]](s32), [[UV1]](s32), [[UV2]](s32)
     ; GCN-NEXT: [[BITCAST:%[0-9]+]]:_(<6 x s16>) = G_BITCAST [[BUILD_VECTOR]](<3 x s32>)
     ; GCN-NEXT: S_ENDPGM 0, implicit [[BITCAST]](<6 x s16>)
     %0:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
     %1:_(s32) = G_CONSTANT i32 0
-    %2:_(<6 x s16>) = G_INTRINSIC intrinsic(@llvm.amdgcn.s.buffer.load), %0, %1, 0
+    %2:_(<6 x s16>) = G_INTRINSIC_W_SIDE_EFFECTS intrinsic(@llvm.amdgcn.s.buffer.load), %0, %1, 0 :: (dereferenceable invariant load (s96), align 1)
     S_ENDPGM 0, implicit %2
 
 ...
 
 ---
 name: s_buffer_load_v6s32
 body:             |
   bb.0:
     liveins: $sgpr0_sgpr1_sgpr2_sgpr3
 
     ; GCN-LABEL: name: s_buffer_load_v6s32
     ; GCN: liveins: $sgpr0_sgpr1_sgpr2_sgpr3
     ; GCN-NEXT: {{  $}}
     ; GCN-NEXT: [[COPY:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
     ; GCN-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
-    ; GCN-NEXT: [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:_(<8 x s32>) = G_AMDGPU_S_BUFFER_LOAD [[COPY]](<4 x s32>), [[C]](s32), 0 :: (dereferenceable invariant load (s192), align 4)
+    ; GCN-NEXT: [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:_(<8 x s32>) = G_AMDGPU_S_BUFFER_LOAD [[COPY]](<4 x s32>), [[C]](s32), 0 :: (dereferenceable invariant load (s192), align 8)
     ; GCN-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_S_BUFFER_LOAD]](<8 x s32>)
     ; GCN-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<6 x s32>) = G_BUILD_VECTOR [[UV]](s32), [[UV1]](s32), [[UV2]](s32), [[UV3]](s32), [[UV4]](s32), [[UV5]](s32)
     ; GCN-NEXT: S_ENDPGM 0, implicit [[BUILD_VECTOR]](<6 x s32>)
     %0:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
     %1:_(s32) = G_CONSTANT i32 0
-    %2:_(<6 x s32>) = G_INTRINSIC intrinsic(@llvm.amdgcn.s.buffer.load), %0, %1, 0
+    %2:_(<6 x s32>) = G_INTRINSIC_W_SIDE_EFFECTS intrinsic(@llvm.amdgcn.s.buffer.load), %0, %1, 0 :: (dereferenceable invariant load (s192), align 8)
     S_ENDPGM 0, implicit %2
 
 ...
 
 ---
 name: s_buffer_load_v3s64
 body:             |
   bb.0:
     liveins: $sgpr0_sgpr1_sgpr2_sgpr3
 
     ; GCN-LABEL: name: s_buffer_load_v3s64
     ; GCN: liveins: $sgpr0_sgpr1_sgpr2_sgpr3
     ; GCN-NEXT: {{  $}}
     ; GCN-NEXT: [[COPY:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
     ; GCN-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
     ; GCN-NEXT: [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:_(<4 x s64>) = G_AMDGPU_S_BUFFER_LOAD [[COPY]](<4 x s32>), [[C]](s32), 0 :: (dereferenceable invariant load (s192), align 4)
     ; GCN-NEXT: [[UV:%[0-9]+]]:_(s64), [[UV1:%[0-9]+]]:_(s64), [[UV2:%[0-9]+]]:_(s64), [[UV3:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[AMDGPU_S_BUFFER_LOAD]](<4 x s64>)
     ; GCN-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s64>) = G_BUILD_VECTOR [[UV]](s64), [[UV1]](s64), [[UV2]](s64)
     ; GCN-NEXT: S_ENDPGM 0, implicit [[BUILD_VECTOR]](<3 x s64>)
     %0:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
     %1:_(s32) = G_CONSTANT i32 0
-    %2:_(<3 x s64>) = G_INTRINSIC intrinsic(@llvm.amdgcn.s.buffer.load), %0, %1, 0
+    %2:_(<3 x s64>) = G_INTRINSIC_W_SIDE_EFFECTS intrinsic(@llvm.amdgcn.s.buffer.load), %0, %1, 0 :: (dereferenceable invariant load (s192), align 4)
     S_ENDPGM 0, implicit %2
 
 ...
 
 ---
 name: s_buffer_load_v12s8
 body:             |
   bb.0:
bb.0:
     ; GCN-NEXT: [[AND11:%[0-9]+]]:_(s32) = G_AND [[LSHR8]], [[C4]]
     ; GCN-NEXT: [[SHL5:%[0-9]+]]:_(s32) = G_SHL [[AND11]], [[C2]](s32)
     ; GCN-NEXT: [[OR5:%[0-9]+]]:_(s32) = G_OR [[AND10]], [[SHL5]]
     ; GCN-NEXT: [[BITCAST5:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR5]](s32)
     ; GCN-NEXT: [[CONCAT_VECTORS:%[0-9]+]]:_(<12 x s16>) = G_CONCAT_VECTORS [[BITCAST]](<2 x s16>), [[BITCAST1]](<2 x s16>), [[BITCAST2]](<2 x s16>), [[BITCAST3]](<2 x s16>), [[BITCAST4]](<2 x s16>), [[BITCAST5]](<2 x s16>)
     ; GCN-NEXT: S_ENDPGM 0, implicit [[CONCAT_VECTORS]](<12 x s16>)
     %0:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
     %1:_(s32) = G_CONSTANT i32 0
-    %2:_(<12 x s8>) = G_INTRINSIC intrinsic(@llvm.amdgcn.s.buffer.load), %0, %1, 0
+    %2:_(<12 x s8>) = G_INTRINSIC_W_SIDE_EFFECTS intrinsic(@llvm.amdgcn.s.buffer.load), %0, %1, 0 :: (dereferenceable invariant load (s96), align 4)
     %3:_(<12 x s16>) = G_ANYEXT %2
     S_ENDPGM 0, implicit %3
 
 ...
 
 ---
 name: s_buffer_load_s96
 body:             |
   bb.0:
     liveins: $sgpr0_sgpr1_sgpr2_sgpr3
 
     ; GCN-LABEL: name: s_buffer_load_s96
     ; GCN: liveins: $sgpr0_sgpr1_sgpr2_sgpr3
     ; GCN-NEXT: {{  $}}
     ; GCN-NEXT: [[COPY:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
     ; GCN-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
     ; GCN-NEXT: [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:_(<4 x s32>) = G_AMDGPU_S_BUFFER_LOAD [[COPY]](<4 x s32>), [[C]](s32), 0 :: (dereferenceable invariant load (s96), align 4)
     ; GCN-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[AMDGPU_S_BUFFER_LOAD]](<4 x s32>)
     ; GCN-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[UV]](s32), [[UV1]](s32), [[UV2]](s32)
     ; GCN-NEXT: S_ENDPGM 0, implicit [[BUILD_VECTOR]](<3 x s32>)
     %0:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
     %1:_(s32) = G_CONSTANT i32 0
-    %2:_(<3 x s32>) = G_INTRINSIC intrinsic(@llvm.amdgcn.s.buffer.load), %0, %1, 0
+    %2:_(<3 x s32>) = G_INTRINSIC_W_SIDE_EFFECTS intrinsic(@llvm.amdgcn.s.buffer.load), %0, %1, 0 :: (dereferenceable invariant load (s96), align 4)
     S_ENDPGM 0, implicit %2
 
 ...

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.s.buffer.load.ll

define amdgpu_ps i32 @s_buffer_load_i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
   ; GFX6-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX7-LABEL: name: s_buffer_load_i32
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
   ; GFX7-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX8-LABEL: name: s_buffer_load_i32
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
   ; GFX8-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   %val = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %rsrc, i32 %soffset, i32 0)
   ret i32 %val
 }
 
 define amdgpu_ps i32 @s_buffer_load_i32_glc(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
   ; GFX6-LABEL: name: s_buffer_load_i32_glc
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[COPY4]], 1 :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[COPY4]], 1 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
   ; GFX6-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX7-LABEL: name: s_buffer_load_i32_glc
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[COPY4]], 1 :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[COPY4]], 1 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
   ; GFX7-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX8-LABEL: name: s_buffer_load_i32_glc
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[COPY4]], 1 :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[COPY4]], 1 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
   ; GFX8-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   %val = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %rsrc, i32 %soffset, i32 1)
   ret i32 %val
 }
 
 define amdgpu_ps <2 x i32> @s_buffer_load_v2i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
   ; GFX6-LABEL: name: s_buffer_load_v2i32
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORDX2_SGPR:%[0-9]+]]:sreg_64_xexec = S_BUFFER_LOAD_DWORDX2_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s64), align 4)
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORDX2_SGPR:%[0-9]+]]:sreg_64_xexec = S_BUFFER_LOAD_DWORDX2_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (<2 x s32>), align 4, addrspace 7)
   ; GFX6-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX2_SGPR]].sub0
   ; GFX6-NEXT:   [[COPY6:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX2_SGPR]].sub1
   ; GFX6-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
   ; GFX6-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX6-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY6]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
   ; GFX6-NEXT:   $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1
   ; GFX7-LABEL: name: s_buffer_load_v2i32
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORDX2_SGPR:%[0-9]+]]:sreg_64_xexec = S_BUFFER_LOAD_DWORDX2_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s64), align 4)
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORDX2_SGPR:%[0-9]+]]:sreg_64_xexec = S_BUFFER_LOAD_DWORDX2_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (<2 x s32>), align 4, addrspace 7)
   ; GFX7-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX2_SGPR]].sub0
   ; GFX7-NEXT:   [[COPY6:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX2_SGPR]].sub1
   ; GFX7-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
   ; GFX7-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX7-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY6]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
   ; GFX7-NEXT:   $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1
   ; GFX8-LABEL: name: s_buffer_load_v2i32
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORDX2_SGPR:%[0-9]+]]:sreg_64_xexec = S_BUFFER_LOAD_DWORDX2_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s64), align 4)
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORDX2_SGPR:%[0-9]+]]:sreg_64_xexec = S_BUFFER_LOAD_DWORDX2_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (<2 x s32>), align 4, addrspace 7)
   ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX2_SGPR]].sub0
   ; GFX8-NEXT:   [[COPY6:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX2_SGPR]].sub1
   ; GFX8-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
   ; GFX8-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX8-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY6]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
   ; GFX8-NEXT:   $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1
   %val = call <2 x i32> @llvm.amdgcn.s.buffer.load.v2i32(<4 x i32> %rsrc, i32 %soffset, i32 0)
   ret <2 x i32> %val
 }
 
 define amdgpu_ps <3 x i32> @s_buffer_load_v3i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
   ; GFX6-LABEL: name: s_buffer_load_v3i32
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORDX4_SGPR:%[0-9]+]]:sgpr_128 = S_BUFFER_LOAD_DWORDX4_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s96), align 4)
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORDX4_SGPR:%[0-9]+]]:sgpr_128 = S_BUFFER_LOAD_DWORDX4_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (<3 x s32>), align 4, addrspace 7)
   ; GFX6-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX4_SGPR]].sub0
   ; GFX6-NEXT:   [[COPY6:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX4_SGPR]].sub1
   ; GFX6-NEXT:   [[COPY7:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX4_SGPR]].sub2
   ; GFX6-NEXT:   [[COPY8:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX4_SGPR]].sub3
   ; GFX6-NEXT:   [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY9]], implicit $exec
   ; GFX6-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX6-NEXT:   [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[COPY6]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY10]], implicit $exec
   ; GFX6-NEXT:   $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
   ; GFX6-NEXT:   [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[COPY7]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY11]], implicit $exec
   ; GFX6-NEXT:   $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1, implicit $sgpr2
   ; GFX7-LABEL: name: s_buffer_load_v3i32
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORDX4_SGPR:%[0-9]+]]:sgpr_128 = S_BUFFER_LOAD_DWORDX4_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s96), align 4)
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORDX4_SGPR:%[0-9]+]]:sgpr_128 = S_BUFFER_LOAD_DWORDX4_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (<3 x s32>), align 4, addrspace 7)
   ; GFX7-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX4_SGPR]].sub0
   ; GFX7-NEXT:   [[COPY6:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX4_SGPR]].sub1
   ; GFX7-NEXT:   [[COPY7:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX4_SGPR]].sub2
   ; GFX7-NEXT:   [[COPY8:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX4_SGPR]].sub3
   ; GFX7-NEXT:   [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY9]], implicit $exec
   ; GFX7-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX7-NEXT:   [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[COPY6]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY10]], implicit $exec
   ; GFX7-NEXT:   $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
   ; GFX7-NEXT:   [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[COPY7]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY11]], implicit $exec
   ; GFX7-NEXT:   $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1, implicit $sgpr2
   ; GFX8-LABEL: name: s_buffer_load_v3i32
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORDX4_SGPR:%[0-9]+]]:sgpr_128 = S_BUFFER_LOAD_DWORDX4_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s96), align 4)
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORDX4_SGPR:%[0-9]+]]:sgpr_128 = S_BUFFER_LOAD_DWORDX4_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (<3 x s32>), align 4, addrspace 7)
   ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX4_SGPR]].sub0
   ; GFX8-NEXT:   [[COPY6:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX4_SGPR]].sub1
   ; GFX8-NEXT:   [[COPY7:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX4_SGPR]].sub2
   ; GFX8-NEXT:   [[COPY8:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX4_SGPR]].sub3
   ; GFX8-NEXT:   [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY9]], implicit $exec
   ; GFX8-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX8-NEXT:   [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[COPY6]]
define amdgpu_ps <8 x i32> @s_buffer_load_v8i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORDX8_SGPR:%[0-9]+]]:sgpr_256 = S_BUFFER_LOAD_DWORDX8_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s256), align 4)
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORDX8_SGPR:%[0-9]+]]:sgpr_256 = S_BUFFER_LOAD_DWORDX8_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (<8 x s32>), align 4, addrspace 7)
   ; GFX6-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub0
   ; GFX6-NEXT:   [[COPY6:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub1
   ; GFX6-NEXT:   [[COPY7:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub2
   ; GFX6-NEXT:   [[COPY8:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub3
   ; GFX6-NEXT:   [[COPY9:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub4
   ; GFX6-NEXT:   [[COPY10:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub5
   ; GFX6-NEXT:   [[COPY11:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub6
   ; GFX6-NEXT:   [[COPY12:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub7
define amdgpu_ps <8 x i32> @s_buffer_load_v8i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORDX8_SGPR:%[0-9]+]]:sgpr_256 = S_BUFFER_LOAD_DWORDX8_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s256), align 4)
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORDX8_SGPR:%[0-9]+]]:sgpr_256 = S_BUFFER_LOAD_DWORDX8_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (<8 x s32>), align 4, addrspace 7)
   ; GFX7-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub0
   ; GFX7-NEXT:   [[COPY6:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub1
   ; GFX7-NEXT:   [[COPY7:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub2
   ; GFX7-NEXT:   [[COPY8:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub3
   ; GFX7-NEXT:   [[COPY9:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub4
   ; GFX7-NEXT:   [[COPY10:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub5
   ; GFX7-NEXT:   [[COPY11:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub6
   ; GFX7-NEXT:   [[COPY12:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub7
define amdgpu_ps <8 x i32> @s_buffer_load_v8i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORDX8_SGPR:%[0-9]+]]:sgpr_256 = S_BUFFER_LOAD_DWORDX8_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s256), align 4)
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORDX8_SGPR:%[0-9]+]]:sgpr_256 = S_BUFFER_LOAD_DWORDX8_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (<8 x s32>), align 4, addrspace 7)
   ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub0
   ; GFX8-NEXT:   [[COPY6:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub1
   ; GFX8-NEXT:   [[COPY7:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub2
   ; GFX8-NEXT:   [[COPY8:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub3
   ; GFX8-NEXT:   [[COPY9:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub4
   ; GFX8-NEXT:   [[COPY10:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub5
   ; GFX8-NEXT:   [[COPY11:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub6
   ; GFX8-NEXT:   [[COPY12:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX8_SGPR]].sub7
define amdgpu_ps <16 x i32> @s_buffer_load_v16i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORDX16_SGPR:%[0-9]+]]:sgpr_512 = S_BUFFER_LOAD_DWORDX16_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s512), align 4)
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORDX16_SGPR:%[0-9]+]]:sgpr_512 = S_BUFFER_LOAD_DWORDX16_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (<16 x s32>), align 4, addrspace 7)
   ; GFX6-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub0
   ; GFX6-NEXT:   [[COPY6:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub1
   ; GFX6-NEXT:   [[COPY7:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub2
   ; GFX6-NEXT:   [[COPY8:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub3
   ; GFX6-NEXT:   [[COPY9:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub4
   ; GFX6-NEXT:   [[COPY10:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub5
   ; GFX6-NEXT:   [[COPY11:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub6
   ; GFX6-NEXT:   [[COPY12:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub7
define amdgpu_ps <16 x i32> @s_buffer_load_v16i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORDX16_SGPR:%[0-9]+]]:sgpr_512 = S_BUFFER_LOAD_DWORDX16_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s512), align 4)
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORDX16_SGPR:%[0-9]+]]:sgpr_512 = S_BUFFER_LOAD_DWORDX16_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (<16 x s32>), align 4, addrspace 7)
   ; GFX7-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub0
   ; GFX7-NEXT:   [[COPY6:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub1
   ; GFX7-NEXT:   [[COPY7:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub2
   ; GFX7-NEXT:   [[COPY8:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub3
   ; GFX7-NEXT:   [[COPY9:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub4
   ; GFX7-NEXT:   [[COPY10:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub5
   ; GFX7-NEXT:   [[COPY11:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub6
   ; GFX7-NEXT:   [[COPY12:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub7
define amdgpu_ps <16 x i32> @s_buffer_load_v16i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORDX16_SGPR:%[0-9]+]]:sgpr_512 = S_BUFFER_LOAD_DWORDX16_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s512), align 4)
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORDX16_SGPR:%[0-9]+]]:sgpr_512 = S_BUFFER_LOAD_DWORDX16_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (<16 x s32>), align 4, addrspace 7)
   ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub0
   ; GFX8-NEXT:   [[COPY6:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub1
   ; GFX8-NEXT:   [[COPY7:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub2
   ; GFX8-NEXT:   [[COPY8:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub3
   ; GFX8-NEXT:   [[COPY9:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub4
   ; GFX8-NEXT:   [[COPY10:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub5
   ; GFX8-NEXT:   [[COPY11:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub6
   ; GFX8-NEXT:   [[COPY12:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX16_SGPR]].sub7
define amdgpu_ps i32 @s_buffer_load_i32_offset_1(<4 x i32> inreg %rsrc) {
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX6-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX7-LABEL: name: s_buffer_load_i32_offset_1
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX7-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX8-LABEL: name: s_buffer_load_i32_offset_1
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 1, 0 :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 1, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX8-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   %val = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %rsrc, i32 1, i32 0)
   ret i32 %val
 }
 
 define amdgpu_ps i32 @s_buffer_load_i32_offset_glc_4(<4 x i32> inreg %rsrc) {
   ; GFX6-LABEL: name: s_buffer_load_i32_offset_glc_4
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 1, 1 :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 1, 1 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX6-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX7-LABEL: name: s_buffer_load_i32_offset_glc_4
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 1, 1 :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 1, 1 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX7-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX8-LABEL: name: s_buffer_load_i32_offset_glc_4
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 4, 1 :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 4, 1 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX8-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   %val = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %rsrc, i32 4, i32 1)
   ret i32 %val
 }
 
 define amdgpu_ps i32 @s_buffer_load_i32_offset_255(<4 x i32> inreg %rsrc) {
   ; GFX6-LABEL: name: s_buffer_load_i32_offset_255
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 255
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX6-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX7-LABEL: name: s_buffer_load_i32_offset_255
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 255
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX7-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX8-LABEL: name: s_buffer_load_i32_offset_255
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 255, 0 :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 255, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX8-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   %val = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %rsrc, i32 255, i32 0)
   ret i32 %val
 }
 
 define amdgpu_ps i32 @s_buffer_load_i32_offset_256(<4 x i32> inreg %rsrc) {
   ; GFX6-LABEL: name: s_buffer_load_i32_offset_256
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 64, 0 :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 64, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX6-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX7-LABEL: name: s_buffer_load_i32_offset_256
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 64, 0 :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 64, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX7-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX8-LABEL: name: s_buffer_load_i32_offset_256
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 256, 0 :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 256, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX8-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   %val = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %rsrc, i32 256, i32 0)
   ret i32 %val
 }
 
 define amdgpu_ps i32 @s_buffer_load_i32_offset_1020(<4 x i32> inreg %rsrc) {
   ; GFX6-LABEL: name: s_buffer_load_i32_offset_1020
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 255, 0 :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 255, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX6-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX7-LABEL: name: s_buffer_load_i32_offset_1020
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 255, 0 :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 255, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX7-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX8-LABEL: name: s_buffer_load_i32_offset_1020
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 1020, 0 :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 1020, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX8-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   %val = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %rsrc, i32 1020, i32 0)
   ret i32 %val
 }
 
 define amdgpu_ps i32 @s_buffer_load_i32_offset_1023(<4 x i32> inreg %rsrc) {
   ; GFX6-LABEL: name: s_buffer_load_i32_offset_1023
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1023
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX6-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX7-LABEL: name: s_buffer_load_i32_offset_1023
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1023
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX7-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX8-LABEL: name: s_buffer_load_i32_offset_1023
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 1023, 0 :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 1023, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX8-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   %val = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %rsrc, i32 1023, i32 0)
   ret i32 %val
 }
 
 define amdgpu_ps i32 @s_buffer_load_i32_offset_1024(<4 x i32> inreg %rsrc) {
   ; GFX6-LABEL: name: s_buffer_load_i32_offset_1024
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1024
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX6-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX7-LABEL: name: s_buffer_load_i32_offset_1024
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 256, 0 :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 256, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX7-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX8-LABEL: name: s_buffer_load_i32_offset_1024
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 1024, 0 :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 1024, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX8-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   %val = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %rsrc, i32 1024, i32 0)
   ret i32 %val
 }
 
 define amdgpu_ps i32 @s_buffer_load_i32_offset_1025(<4 x i32> inreg %rsrc) {
   ; GFX6-LABEL: name: s_buffer_load_i32_offset_1025
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1025
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX6-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX7-LABEL: name: s_buffer_load_i32_offset_1025
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1025
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX7-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX8-LABEL: name: s_buffer_load_i32_offset_1025
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 1025, 0 :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 1025, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX8-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   %val = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %rsrc, i32 1025, i32 0)
   ret i32 %val
 }
 
 define amdgpu_ps i32 @s_buffer_load_i32_offset_neg1(<4 x i32> inreg %desc) {
   ; GFX6-LABEL: name: s_buffer_load_i32_offset_neg1
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX6-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX7-LABEL: name: s_buffer_load_i32_offset_neg1
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX7-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX8-LABEL: name: s_buffer_load_i32_offset_neg1
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX8-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   %load = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %desc, i32 -1, i32 0)
   ret i32 %load
 }
 
 define amdgpu_ps i32 @s_buffer_load_i32_offset_neg4(<4 x i32> inreg %desc) {
   ; GFX6-LABEL: name: s_buffer_load_i32_offset_neg4
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -4
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX6-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX7-LABEL: name: s_buffer_load_i32_offset_neg4
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 1073741823, 0 :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 1073741823, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX7-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX8-LABEL: name: s_buffer_load_i32_offset_neg4
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -4
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX8-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   %load = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %desc, i32 -4, i32 0)
   ret i32 %load
 }
 
 define amdgpu_ps i32 @s_buffer_load_i32_offset_neg8(<4 x i32> inreg %desc) {
   ; GFX6-LABEL: name: s_buffer_load_i32_offset_neg8
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -8
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX6-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX7-LABEL: name: s_buffer_load_i32_offset_neg8
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 1073741822, 0 :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 1073741822, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX7-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX8-LABEL: name: s_buffer_load_i32_offset_neg8
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -8
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX8-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   %load = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %desc, i32 -8, i32 0)
   ret i32 %load
 }
 
 define amdgpu_ps i32 @s_buffer_load_i32_offset_bit31(<4 x i32> inreg %desc) {
   ; GFX6-LABEL: name: s_buffer_load_i32_offset_bit31
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -2147483648
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX6-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX7-LABEL: name: s_buffer_load_i32_offset_bit31
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 536870912, 0 :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 536870912, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX7-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX8-LABEL: name: s_buffer_load_i32_offset_bit31
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -2147483648
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX8-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   %load = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %desc, i32 -2147483648, i32 0)
   ret i32 %load
 }
 
 define amdgpu_ps i32 @s_buffer_load_i32_offset_glc_bit30(<4 x i32> inreg %desc) {
   ; GFX6-LABEL: name: s_buffer_load_i32_offset_glc_bit30
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1073741824
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 1 :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 1 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX6-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX7-LABEL: name: s_buffer_load_i32_offset_glc_bit30
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 268435456, 1 :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 268435456, 1 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX7-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX8-LABEL: name: s_buffer_load_i32_offset_glc_bit30
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1073741824
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 1 :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 1 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX8-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   %load = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %desc, i32 1073741824, i32 1)
   ret i32 %load
 }
 
 define amdgpu_ps i32 @s_buffer_load_i32_offset_bit29(<4 x i32> inreg %desc) {
   ; GFX6-LABEL: name: s_buffer_load_i32_offset_bit29
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 536870912
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX6-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX7-LABEL: name: s_buffer_load_i32_offset_bit29
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 134217728, 0 :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 134217728, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX7-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX8-LABEL: name: s_buffer_load_i32_offset_bit29
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 536870912
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX8-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   %load = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %desc, i32 536870912, i32 0)
   ret i32 %load
 }
 
 define amdgpu_ps i32 @s_buffer_load_i32_offset_bit21(<4 x i32> inreg %desc) {
   ; GFX6-LABEL: name: s_buffer_load_i32_offset_bit21
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 2097152
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX6-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX7-LABEL: name: s_buffer_load_i32_offset_bit21
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 524288, 0 :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 524288, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX7-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX8-LABEL: name: s_buffer_load_i32_offset_bit21
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 2097152
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX8-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   %load = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %desc, i32 2097152, i32 0)
   ret i32 %load
 }
 
 define amdgpu_ps i32 @s_buffer_load_i32_offset_bit20(<4 x i32> inreg %desc) {
   ; GFX6-LABEL: name: s_buffer_load_i32_offset_bit20
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1048576
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX6-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX7-LABEL: name: s_buffer_load_i32_offset_bit20
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 262144, 0 :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 262144, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX7-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX8-LABEL: name: s_buffer_load_i32_offset_bit20
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1048576
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX8-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   %load = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %desc, i32 1048576, i32 0)
   ret i32 %load
 }
 
 define amdgpu_ps i32 @s_buffer_load_i32_offset_neg_bit20(<4 x i32> inreg %desc) {
   ; GFX6-LABEL: name: s_buffer_load_i32_offset_neg_bit20
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1048576
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX6-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX7-LABEL: name: s_buffer_load_i32_offset_neg_bit20
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 1073479680, 0 :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 1073479680, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX7-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX8-LABEL: name: s_buffer_load_i32_offset_neg_bit20
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1048576
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX8-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   %load = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %desc, i32  -1048576, i32 0)
   ret i32 %load
 }
 
 define amdgpu_ps i32 @s_buffer_load_i32_offset_bit19(<4 x i32> inreg %desc) {
   ; GFX6-LABEL: name: s_buffer_load_i32_offset_bit19
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 524288
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX6-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX7-LABEL: name: s_buffer_load_i32_offset_bit19
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 131072, 0 :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 131072, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX7-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX8-LABEL: name: s_buffer_load_i32_offset_bit19
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 524288, 0 :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 524288, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX8-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   %load = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %desc, i32 524288, i32 0)
   ret i32 %load
 }
 
 define amdgpu_ps i32 @s_buffer_load_i32_offset_neg_bit19(<4 x i32> inreg %desc) {
   ; GFX6-LABEL: name: s_buffer_load_i32_offset_neg_bit19
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -524288
-  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX6-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX6-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX7-LABEL: name: s_buffer_load_i32_offset_neg_bit19
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
-  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 1073610752, 0 :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 1073610752, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]]
   ; GFX7-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX7-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GFX8-LABEL: name: s_buffer_load_i32_offset_neg_bit19
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -524288
-  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
   ; GFX8-NEXT:   [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
   ; GFX8-NEXT:   $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   %load = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %desc, i32 -524288, i32 0)
   ret i32 %load
 }
 
 ; Check cases that need to be converted to MUBUF due to the offset being a VGPR.
 define amdgpu_ps float @s_buffer_load_f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; GFX6-LABEL: name: s_buffer_load_f32_vgpr_offset
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GFX7-LABEL: name: s_buffer_load_f32_vgpr_offset
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX7-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GFX8-LABEL: name: s_buffer_load_f32_vgpr_offset
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   %val = call float @llvm.amdgcn.s.buffer.load.f32(<4 x i32> %rsrc, i32 %soffset, i32 0)
   ret float %val
 }
 
 define amdgpu_ps <2 x float> @s_buffer_load_v2f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; GFX6-LABEL: name: s_buffer_load_v2f32_vgpr_offset
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX2_OFFEN:%[0-9]+]]:vreg_64 = BUFFER_LOAD_DWORDX2_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s64), align 4)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX2_OFFEN:%[0-9]+]]:vreg_64 = BUFFER_LOAD_DWORDX2_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s64), align 4, addrspace 7)
   ; GFX6-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX2_OFFEN]].sub0
   ; GFX6-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX2_OFFEN]].sub1
   ; GFX6-NEXT:   $vgpr0 = COPY [[COPY5]]
   ; GFX6-NEXT:   $vgpr1 = COPY [[COPY6]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1
   ; GFX7-LABEL: name: s_buffer_load_v2f32_vgpr_offset
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX7-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX2_OFFEN:%[0-9]+]]:vreg_64 = BUFFER_LOAD_DWORDX2_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s64), align 4)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX2_OFFEN:%[0-9]+]]:vreg_64 = BUFFER_LOAD_DWORDX2_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s64), align 4, addrspace 7)
   ; GFX7-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX2_OFFEN]].sub0
   ; GFX7-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX2_OFFEN]].sub1
   ; GFX7-NEXT:   $vgpr0 = COPY [[COPY5]]
   ; GFX7-NEXT:   $vgpr1 = COPY [[COPY6]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1
   ; GFX8-LABEL: name: s_buffer_load_v2f32_vgpr_offset
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX2_OFFEN:%[0-9]+]]:vreg_64 = BUFFER_LOAD_DWORDX2_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s64), align 4)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX2_OFFEN:%[0-9]+]]:vreg_64 = BUFFER_LOAD_DWORDX2_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s64), align 4, addrspace 7)
   ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX2_OFFEN]].sub0
   ; GFX8-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX2_OFFEN]].sub1
   ; GFX8-NEXT:   $vgpr0 = COPY [[COPY5]]
   ; GFX8-NEXT:   $vgpr1 = COPY [[COPY6]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1
   %val = call <2 x float> @llvm.amdgcn.s.buffer.load.v2f32(<4 x i32> %rsrc, i32 %soffset, i32 0)
   ret <2 x float> %val
 }
 
 define amdgpu_ps <3 x float> @s_buffer_load_v3f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; GFX6-LABEL: name: s_buffer_load_v3f32_vgpr_offset
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX6-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub0
   ; GFX6-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub1
   ; GFX6-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub2
   ; GFX6-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub3
   ; GFX6-NEXT:   $vgpr0 = COPY [[COPY5]]
   ; GFX6-NEXT:   $vgpr1 = COPY [[COPY6]]
   ; GFX6-NEXT:   $vgpr2 = COPY [[COPY7]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
   ; GFX7-LABEL: name: s_buffer_load_v3f32_vgpr_offset
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX7-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX7-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub0
   ; GFX7-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub1
   ; GFX7-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub2
   ; GFX7-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub3
   ; GFX7-NEXT:   $vgpr0 = COPY [[COPY5]]
   ; GFX7-NEXT:   $vgpr1 = COPY [[COPY6]]
   ; GFX7-NEXT:   $vgpr2 = COPY [[COPY7]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
   ; GFX8-LABEL: name: s_buffer_load_v3f32_vgpr_offset
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub0
   ; GFX8-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub1
   ; GFX8-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub2
   ; GFX8-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub3
   ; GFX8-NEXT:   $vgpr0 = COPY [[COPY5]]
   ; GFX8-NEXT:   $vgpr1 = COPY [[COPY6]]
   ; GFX8-NEXT:   $vgpr2 = COPY [[COPY7]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
   %val = call <3 x float> @llvm.amdgcn.s.buffer.load.v3f32(<4 x i32> %rsrc, i32 %soffset, i32 0)
   ret <3 x float> %val
 }
 
 define amdgpu_ps <4 x float> @s_buffer_load_v4f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; GFX6-LABEL: name: s_buffer_load_v4f32_vgpr_offset
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX6-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub0
   ; GFX6-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub1
   ; GFX6-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub2
   ; GFX6-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub3
   ; GFX6-NEXT:   $vgpr0 = COPY [[COPY5]]
   ; GFX6-NEXT:   $vgpr1 = COPY [[COPY6]]
   ; GFX6-NEXT:   $vgpr2 = COPY [[COPY7]]
   ; GFX6-NEXT:   $vgpr3 = COPY [[COPY8]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
   ; GFX7-LABEL: name: s_buffer_load_v4f32_vgpr_offset
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX7-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX7-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub0
   ; GFX7-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub1
   ; GFX7-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub2
   ; GFX7-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub3
   ; GFX7-NEXT:   $vgpr0 = COPY [[COPY5]]
   ; GFX7-NEXT:   $vgpr1 = COPY [[COPY6]]
   ; GFX7-NEXT:   $vgpr2 = COPY [[COPY7]]
   ; GFX7-NEXT:   $vgpr3 = COPY [[COPY8]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
   ; GFX8-LABEL: name: s_buffer_load_v4f32_vgpr_offset
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub0
   ; GFX8-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub1
   ; GFX8-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub2
   ; GFX8-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub3
   ; GFX8-NEXT:   $vgpr0 = COPY [[COPY5]]
   ; GFX8-NEXT:   $vgpr1 = COPY [[COPY6]]
   ; GFX8-NEXT:   $vgpr2 = COPY [[COPY7]]
   ; GFX8-NEXT:   $vgpr3 = COPY [[COPY8]]
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX6-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_256 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7
   ; GFX6-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
   ; GFX6-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
   ; GFX6-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub2
   ; GFX6-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub3
   ; GFX6-NEXT:   [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub4
   ; GFX6-NEXT:   [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub5
   ; GFX6-NEXT:   [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub6
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX7-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX7-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_256 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7
   ; GFX7-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
   ; GFX7-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
   ; GFX7-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub2
   ; GFX7-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub3
   ; GFX7-NEXT:   [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub4
   ; GFX7-NEXT:   [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub5
   ; GFX7-NEXT:   [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub6
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX8-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_256 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7
   ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
   ; GFX8-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
   ; GFX8-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub2
   ; GFX8-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub3
   ; GFX8-NEXT:   [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub4
   ; GFX8-NEXT:   [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub5
   ; GFX8-NEXT:   [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub6
define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 32, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4)
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 48, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 32, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4, addrspace 7)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 48, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4, addrspace 7)
   ; GFX6-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_512 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7, [[BUFFER_LOAD_DWORDX4_OFFEN2]], %subreg.sub8_sub9_sub10_sub11, [[BUFFER_LOAD_DWORDX4_OFFEN3]], %subreg.sub12_sub13_sub14_sub15
   ; GFX6-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
   ; GFX6-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
   ; GFX6-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub2
   ; GFX6-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub3
   ; GFX6-NEXT:   [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub4
   ; GFX6-NEXT:   [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub5
   ; GFX6-NEXT:   [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub6
define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX7-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 32, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4)
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 48, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 32, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4, addrspace 7)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 48, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4, addrspace 7)
   ; GFX7-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_512 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7, [[BUFFER_LOAD_DWORDX4_OFFEN2]], %subreg.sub8_sub9_sub10_sub11, [[BUFFER_LOAD_DWORDX4_OFFEN3]], %subreg.sub12_sub13_sub14_sub15
   ; GFX7-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
   ; GFX7-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
   ; GFX7-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub2
   ; GFX7-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub3
   ; GFX7-NEXT:   [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub4
   ; GFX7-NEXT:   [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub5
   ; GFX7-NEXT:   [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub6
define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 32, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4)
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 48, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 32, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4, addrspace 7)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 48, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4, addrspace 7)
   ; GFX8-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_512 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7, [[BUFFER_LOAD_DWORDX4_OFFEN2]], %subreg.sub8_sub9_sub10_sub11, [[BUFFER_LOAD_DWORDX4_OFFEN3]], %subreg.sub12_sub13_sub14_sub15
   ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
   ; GFX8-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
   ; GFX8-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub2
   ; GFX8-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub3
   ; GFX8-NEXT:   [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub4
   ; GFX8-NEXT:   [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub5
   ; GFX8-NEXT:   [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub6
define amdgpu_ps float @s_buffer_load_f32_vgpr_offset_add_4092(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4092, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4092, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GFX7-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4092
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX7-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4092, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4092, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GFX8-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4092
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4092, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4092, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   %soffset = add i32 %soffset.base, 4092
   %val = call float @llvm.amdgcn.s.buffer.load.f32(<4 x i32> %rsrc, i32 %soffset, i32 0)
   ret float %val
 }
 
 define amdgpu_ps float @s_buffer_load_f32_vgpr_offset_add_4095(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; GFX6-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4095
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GFX7-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4095
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX7-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GFX8-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4095
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   %soffset = add i32 %soffset.base, 4095
   %val = call float @llvm.amdgcn.s.buffer.load.f32(<4 x i32> %rsrc, i32 %soffset, i32 0)
   ret float %val
 }
 
 define amdgpu_ps float @s_buffer_load_f32_vgpr_offset_add_4096(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; GFX6-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4096
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GFX7-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4096
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX7-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GFX8-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4096
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   %soffset = add i32 %soffset.base, 4096
   %val = call float @llvm.amdgcn.s.buffer.load.f32(<4 x i32> %rsrc, i32 %soffset, i32 0)
   ret float %val
 }
 
 ; Make sure the base offset is added to each split load.
 define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_add_4064(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; GFX6-LABEL: name: s_buffer_load_v8f32_vgpr_offset_add_4064
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX6-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_256 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7
   ; GFX6-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
   ; GFX6-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
   ; GFX6-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub2
   ; GFX6-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub3
   ; GFX6-NEXT:   [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub4
   ; GFX6-NEXT:   [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub5
   ; GFX6-NEXT:   [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub6
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_add_4064(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX7-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX7-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_256 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7
   ; GFX7-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
   ; GFX7-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
   ; GFX7-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub2
   ; GFX7-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub3
   ; GFX7-NEXT:   [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub4
   ; GFX7-NEXT:   [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub5
   ; GFX7-NEXT:   [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub6
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_add_4064(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX8-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_256 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7
   ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
   ; GFX8-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
   ; GFX8-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub2
   ; GFX8-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub3
   ; GFX8-NEXT:   [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub4
   ; GFX8-NEXT:   [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub5
   ; GFX8-NEXT:   [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub6
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_add_4068(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4068
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX6-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_256 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7
   ; GFX6-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
   ; GFX6-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
   ; GFX6-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub2
   ; GFX6-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub3
   ; GFX6-NEXT:   [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub4
   ; GFX6-NEXT:   [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub5
   ; GFX6-NEXT:   [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub6
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_add_4068(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX7-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4068
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX7-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_256 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7
   ; GFX7-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
   ; GFX7-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
   ; GFX7-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub2
   ; GFX7-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub3
   ; GFX7-NEXT:   [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub4
   ; GFX7-NEXT:   [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub5
   ; GFX7-NEXT:   [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub6
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_add_4068(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX8-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_256 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7
   ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
   ; GFX8-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
   ; GFX8-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub2
   ; GFX8-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub3
   ; GFX8-NEXT:   [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub4
   ; GFX8-NEXT:   [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub5
   ; GFX8-NEXT:   [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub6
define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset_add_4032(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4032, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4048, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4)
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4032, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4048, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4, addrspace 7)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4, addrspace 7)
   ; GFX6-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_512 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7, [[BUFFER_LOAD_DWORDX4_OFFEN2]], %subreg.sub8_sub9_sub10_sub11, [[BUFFER_LOAD_DWORDX4_OFFEN3]], %subreg.sub12_sub13_sub14_sub15
   ; GFX6-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
   ; GFX6-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
   ; GFX6-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub2
   ; GFX6-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub3
   ; GFX6-NEXT:   [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub4
   ; GFX6-NEXT:   [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub5
   ; GFX6-NEXT:   [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub6
define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset_add_4032(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX7-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4032, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4048, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4)
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4032, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4048, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4, addrspace 7)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4, addrspace 7)
   ; GFX7-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_512 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7, [[BUFFER_LOAD_DWORDX4_OFFEN2]], %subreg.sub8_sub9_sub10_sub11, [[BUFFER_LOAD_DWORDX4_OFFEN3]], %subreg.sub12_sub13_sub14_sub15
   ; GFX7-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
   ; GFX7-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
   ; GFX7-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub2
   ; GFX7-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub3
   ; GFX7-NEXT:   [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub4
   ; GFX7-NEXT:   [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub5
   ; GFX7-NEXT:   [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub6
define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset_add_4032(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4032, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4048, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4)
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4032, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4048, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4, addrspace 7)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4, addrspace 7)
   ; GFX8-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_512 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7, [[BUFFER_LOAD_DWORDX4_OFFEN2]], %subreg.sub8_sub9_sub10_sub11, [[BUFFER_LOAD_DWORDX4_OFFEN3]], %subreg.sub12_sub13_sub14_sub15
   ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
   ; GFX8-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
   ; GFX8-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub2
   ; GFX8-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub3
   ; GFX8-NEXT:   [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub4
   ; GFX8-NEXT:   [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub5
   ; GFX8-NEXT:   [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub6
define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset_add_4036(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4036
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 32, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4)
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 48, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 32, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4, addrspace 7)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 48, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4, addrspace 7)
   ; GFX6-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_512 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7, [[BUFFER_LOAD_DWORDX4_OFFEN2]], %subreg.sub8_sub9_sub10_sub11, [[BUFFER_LOAD_DWORDX4_OFFEN3]], %subreg.sub12_sub13_sub14_sub15
   ; GFX6-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
   ; GFX6-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
   ; GFX6-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub2
   ; GFX6-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub3
   ; GFX6-NEXT:   [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub4
   ; GFX6-NEXT:   [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub5
   ; GFX6-NEXT:   [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub6
define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset_add_4036(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX7-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4036
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 32, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4)
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 48, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 32, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4, addrspace 7)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 48, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4, addrspace 7)
   ; GFX7-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_512 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7, [[BUFFER_LOAD_DWORDX4_OFFEN2]], %subreg.sub8_sub9_sub10_sub11, [[BUFFER_LOAD_DWORDX4_OFFEN3]], %subreg.sub12_sub13_sub14_sub15
   ; GFX7-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
   ; GFX7-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
   ; GFX7-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub2
   ; GFX7-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub3
   ; GFX7-NEXT:   [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub4
   ; GFX7-NEXT:   [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub5
   ; GFX7-NEXT:   [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub6
define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset_add_4036(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4032, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4048, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4)
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4032, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4048, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4, addrspace 7)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4, addrspace 7)
   ; GFX8-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_512 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7, [[BUFFER_LOAD_DWORDX4_OFFEN2]], %subreg.sub8_sub9_sub10_sub11, [[BUFFER_LOAD_DWORDX4_OFFEN3]], %subreg.sub12_sub13_sub14_sub15
   ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
   ; GFX8-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
   ; GFX8-NEXT:   [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub2
   ; GFX8-NEXT:   [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub3
   ; GFX8-NEXT:   [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub4
   ; GFX8-NEXT:   [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub5
   ; GFX8-NEXT:   [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub6
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc(<4 x i32> %rsrc, i32 inreg %soffset) {
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
   ; GFX6-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX6-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.3:
   ; GFX6-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX6-NEXT: {{  $}}
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX6-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.4:
   ; GFX6-NEXT:   successors: %bb.5(0x80000000)
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX6-NEXT: {{  $}}
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc(<4 x i32> %rsrc, i32 inreg %soffset) {
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
   ; GFX7-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX7-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.3:
   ; GFX7-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX7-NEXT: {{  $}}
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX7-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.4:
   ; GFX7-NEXT:   successors: %bb.5(0x80000000)
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX7-NEXT: {{  $}}
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc(<4 x i32> %rsrc, i32 inreg %soffset) {
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
   ; GFX8-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX8-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.3:
   ; GFX8-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX8-NEXT: {{  $}}
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX8-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.4:
   ; GFX8-NEXT:   successors: %bb.5(0x80000000)
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX8-NEXT: {{  $}}
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_soffset_add_4092(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX6-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX6-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.3:
   ; GFX6-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX6-NEXT: {{  $}}
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4092, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4092, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX6-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.4:
   ; GFX6-NEXT:   successors: %bb.5(0x80000000)
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX6-NEXT: {{  $}}
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_soffset_add_4092(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX7-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX7-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.3:
   ; GFX7-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX7-NEXT: {{  $}}
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4092, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4092, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX7-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.4:
   ; GFX7-NEXT:   successors: %bb.5(0x80000000)
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX7-NEXT: {{  $}}
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_soffset_add_4092(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX8-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX8-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.3:
   ; GFX8-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX8-NEXT: {{  $}}
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4092, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4092, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX8-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.4:
   ; GFX8-NEXT:   successors: %bb.5(0x80000000)
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX8-NEXT: {{  $}}
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_soffset_add_4096(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
   ; GFX6-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX6-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.3:
   ; GFX6-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX6-NEXT: {{  $}}
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX6-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.4:
   ; GFX6-NEXT:   successors: %bb.5(0x80000000)
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX6-NEXT: {{  $}}
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_soffset_add_4096(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
   ; GFX7-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX7-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.3:
   ; GFX7-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX7-NEXT: {{  $}}
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX7-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.4:
   ; GFX7-NEXT:   successors: %bb.5(0x80000000)
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX7-NEXT: {{  $}}
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_soffset_add_4096(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
   ; GFX8-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX8-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.3:
   ; GFX8-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX8-NEXT: {{  $}}
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX8-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.4:
   ; GFX8-NEXT:   successors: %bb.5(0x80000000)
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX8-NEXT: {{  $}}
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_offset_4095(<4 x i32> %rsrc) {
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY6]], [[COPY4]], implicit $exec
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX6-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX6-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.3:
   ; GFX6-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX6-NEXT: {{  $}}
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32) from unknown-address + 4095, align 1)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32) from unknown-address + 4095, align 1, addrspace 7)
   ; GFX6-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX6-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.4:
   ; GFX6-NEXT:   successors: %bb.5(0x80000000)
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX6-NEXT: {{  $}}
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_offset_4095(<4 x i32> %rsrc) {
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY6]], [[COPY4]], implicit $exec
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX7-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX7-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.3:
   ; GFX7-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX7-NEXT: {{  $}}
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32) from unknown-address + 4095, align 1)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32) from unknown-address + 4095, align 1, addrspace 7)
   ; GFX7-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX7-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.4:
   ; GFX7-NEXT:   successors: %bb.5(0x80000000)
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX7-NEXT: {{  $}}
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_offset_4095(<4 x i32> %rsrc) {
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY6]], [[COPY4]], implicit $exec
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX8-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX8-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.3:
   ; GFX8-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX8-NEXT: {{  $}}
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32) from unknown-address + 4095, align 1)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32) from unknown-address + 4095, align 1, addrspace 7)
   ; GFX8-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX8-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.4:
   ; GFX8-NEXT:   successors: %bb.5(0x80000000)
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX8-NEXT: {{  $}}
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_offset_4096(<4 x i32> %rsrc) {
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX6-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX6-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.3:
   ; GFX6-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX6-NEXT: {{  $}}
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX6-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.4:
   ; GFX6-NEXT:   successors: %bb.5(0x80000000)
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX6-NEXT: {{  $}}
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_offset_4096(<4 x i32> %rsrc) {
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX7-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX7-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.3:
   ; GFX7-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX7-NEXT: {{  $}}
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX7-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.4:
   ; GFX7-NEXT:   successors: %bb.5(0x80000000)
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX7-NEXT: {{  $}}
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_offset_4096(<4 x i32> %rsrc) {
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY6]], [[COPY4]], implicit $exec
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX8-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX8-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.3:
   ; GFX8-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX8-NEXT: {{  $}}
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32) from unknown-address + 4096)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32) from unknown-address + 4096, addrspace 7)
   ; GFX8-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX8-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.4:
   ; GFX8-NEXT:   successors: %bb.5(0x80000000)
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX8-NEXT: {{  $}}
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4064(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX6-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX6-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.3:
   ; GFX6-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX6-NEXT: {{  $}}
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX6-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX6-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.4:
   ; GFX6-NEXT:   successors: %bb.5(0x80000000)
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX6-NEXT: {{  $}}
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4064(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX7-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX7-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.3:
   ; GFX7-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX7-NEXT: {{  $}}
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX7-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX7-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.4:
   ; GFX7-NEXT:   successors: %bb.5(0x80000000)
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX7-NEXT: {{  $}}
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4064(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX8-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX8-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.3:
   ; GFX8-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX8-NEXT: {{  $}}
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX8-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX8-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.4:
   ; GFX8-NEXT:   successors: %bb.5(0x80000000)
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX8-NEXT: {{  $}}
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4068(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
   ; GFX6-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX6-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.3:
   ; GFX6-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX6-NEXT: {{  $}}
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX6-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX6-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.4:
   ; GFX6-NEXT:   successors: %bb.5(0x80000000)
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX6-NEXT: {{  $}}
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4068(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
   ; GFX7-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX7-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.3:
   ; GFX7-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX7-NEXT: {{  $}}
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX7-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX7-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.4:
   ; GFX7-NEXT:   successors: %bb.5(0x80000000)
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX7-NEXT: {{  $}}
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4068(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
   ; GFX8-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX8-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.3:
   ; GFX8-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX8-NEXT: {{  $}}
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX8-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX8-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.4:
   ; GFX8-NEXT:   successors: %bb.5(0x80000000)
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX8-NEXT: {{  $}}
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4096(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
   ; GFX6-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX6-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.3:
   ; GFX6-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX6-NEXT: {{  $}}
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX6-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX6-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.4:
   ; GFX6-NEXT:   successors: %bb.5(0x80000000)
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX6-NEXT: {{  $}}
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4096(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
   ; GFX7-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX7-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.3:
   ; GFX7-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX7-NEXT: {{  $}}
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX7-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX7-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.4:
   ; GFX7-NEXT:   successors: %bb.5(0x80000000)
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX7-NEXT: {{  $}}
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4096(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
   ; GFX8-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX8-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.3:
   ; GFX8-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX8-NEXT: {{  $}}
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX8-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX8-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.4:
   ; GFX8-NEXT:   successors: %bb.5(0x80000000)
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX8-NEXT: {{  $}}
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_5000(<4 x i32> %rsrc, i32 %offset.base) {
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX6-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX6-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.3:
   ; GFX6-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX6-NEXT: {{  $}}
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX6-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX6-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.4:
   ; GFX6-NEXT:   successors: %bb.5(0x80000000)
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX6-NEXT: {{  $}}
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_5000(<4 x i32> %rsrc, i32 %offset.base) {
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX7-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX7-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.3:
   ; GFX7-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX7-NEXT: {{  $}}
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX7-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX7-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.4:
   ; GFX7-NEXT:   successors: %bb.5(0x80000000)
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX7-NEXT: {{  $}}
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_5000(<4 x i32> %rsrc, i32 %offset.base) {
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX8-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX8-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.3:
   ; GFX8-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX8-NEXT: {{  $}}
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 936, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 952, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 936, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 952, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX8-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX8-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.4:
   ; GFX8-NEXT:   successors: %bb.5(0x80000000)
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX8-NEXT: {{  $}}
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4076(<4 x i32> %rsrc, i32 %offset.base) {
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX6-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX6-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.3:
   ; GFX6-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX6-NEXT: {{  $}}
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX6-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX6-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.4:
   ; GFX6-NEXT:   successors: %bb.5(0x80000000)
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX6-NEXT: {{  $}}
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4076(<4 x i32> %rsrc, i32 %offset.base) {
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX7-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX7-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.3:
   ; GFX7-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX7-NEXT: {{  $}}
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX7-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX7-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.4:
   ; GFX7-NEXT:   successors: %bb.5(0x80000000)
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX7-NEXT: {{  $}}
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4076(<4 x i32> %rsrc, i32 %offset.base) {
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX8-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX8-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.3:
   ; GFX8-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX8-NEXT: {{  $}}
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX8-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX8-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.4:
   ; GFX8-NEXT:   successors: %bb.5(0x80000000)
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX8-NEXT: {{  $}}
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4080(<4 x i32> %rsrc, i32 %offset.base) {
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX6-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX6-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.3:
   ; GFX6-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX6-NEXT: {{  $}}
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX6-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX6-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.4:
   ; GFX6-NEXT:   successors: %bb.5(0x80000000)
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX6-NEXT: {{  $}}
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4080(<4 x i32> %rsrc, i32 %offset.base) {
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX7-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX7-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.3:
   ; GFX7-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX7-NEXT: {{  $}}
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX7-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX7-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.4:
   ; GFX7-NEXT:   successors: %bb.5(0x80000000)
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX7-NEXT: {{  $}}
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4080(<4 x i32> %rsrc, i32 %offset.base) {
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
   ; GFX8-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX8-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.3:
   ; GFX8-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX8-NEXT: {{  $}}
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GFX8-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX8-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.4:
   ; GFX8-NEXT:   successors: %bb.5(0x80000000)
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX8-NEXT: {{  $}}
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_offset_4064(<4 x i32> %rsrc, i32 %offset.base) {
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY6]], [[COPY4]], implicit $exec
   ; GFX6-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX6-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX6-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.3:
   ; GFX6-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX6-NEXT: {{  $}}
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4)
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4, addrspace 7)
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4, addrspace 7)
   ; GFX6-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX6-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT: bb.4:
   ; GFX6-NEXT:   successors: %bb.5(0x80000000)
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX6-NEXT: {{  $}}
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_offset_4064(<4 x i32> %rsrc, i32 %offset.base) {
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY6]], [[COPY4]], implicit $exec
   ; GFX7-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX7-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX7-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.3:
   ; GFX7-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX7-NEXT: {{  $}}
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4)
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4, addrspace 7)
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4, addrspace 7)
   ; GFX7-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX7-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT: bb.4:
   ; GFX7-NEXT:   successors: %bb.5(0x80000000)
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX7-NEXT: {{  $}}
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_offset_4064(<4 x i32> %rsrc, i32 %offset.base) {
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY6]], [[COPY4]], implicit $exec
   ; GFX8-NEXT:   [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
   ; GFX8-NEXT:   [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
   ; GFX8-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.3:
   ; GFX8-NEXT:   successors: %bb.4(0x40000000), %bb.2(0x40000000)
   ; GFX8-NEXT: {{  $}}
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4)
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4, addrspace 7)
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4, addrspace 7)
   ; GFX8-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GFX8-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT: bb.4:
   ; GFX8-NEXT:   successors: %bb.5(0x80000000)
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GFX8-NEXT: {{  $}}
define amdgpu_ps float @s_buffer_load_f32_offset_add_vgpr_sgpr(<4 x i32> inreg %rsrc, i32 %offset.v, i32 inreg %offset.s) {
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX6-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GFX7-LABEL: name: s_buffer_load_f32_offset_add_vgpr_sgpr
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX7-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GFX8-LABEL: name: s_buffer_load_f32_offset_add_vgpr_sgpr
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   %offset = add i32 %offset.v, %offset.s
   %val = call float @llvm.amdgcn.s.buffer.load.f32(<4 x i32> %rsrc, i32 %offset, i32 0)
   ret float %val
 }
 
 define amdgpu_ps float @s_buffer_load_f32_offset_add_sgpr_vgpr(<4 x i32> inreg %rsrc, i32 %offset.v, i32 inreg %offset.s) {
   ; GFX6-LABEL: name: s_buffer_load_f32_offset_add_sgpr_vgpr
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX6-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GFX7-LABEL: name: s_buffer_load_f32_offset_add_sgpr_vgpr
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX7-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GFX8-LABEL: name: s_buffer_load_f32_offset_add_sgpr_vgpr
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   %offset = add i32 %offset.s, %offset.v
   %val = call float @llvm.amdgcn.s.buffer.load.f32(<4 x i32> %rsrc, i32 %offset, i32 0)
   ret float %val
 }
 
 define amdgpu_ps float @s_buffer_load_f32_offset_add_vgpr_sgpr_imm(<4 x i32> inreg %rsrc, i32 %offset.v, i32 inreg %offset.s) {
   ; GFX6-LABEL: name: s_buffer_load_f32_offset_add_vgpr_sgpr_imm
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX6-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
   ; GFX6-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
   ; GFX6-NEXT:   [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[COPY4]], [[COPY6]], 0, implicit $exec
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GFX7-LABEL: name: s_buffer_load_f32_offset_add_vgpr_sgpr_imm
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX7-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
   ; GFX7-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
   ; GFX7-NEXT:   [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[COPY4]], [[COPY6]], 0, implicit $exec
   ; GFX7-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GFX8-LABEL: name: s_buffer_load_f32_offset_add_vgpr_sgpr_imm
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
   ; GFX8-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
   ; GFX8-NEXT:   [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[COPY4]], [[COPY6]], 0, implicit $exec
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   %offset.base = add i32 %offset.v, %offset.s
   %offset = add i32 %offset.base, 1024
   %val = call float @llvm.amdgcn.s.buffer.load.f32(<4 x i32> %rsrc, i32 %offset, i32 0)
   ret float %val
 }
 
 define amdgpu_ps float @s_buffer_load_f32_offset_add_sgpr_vgpr_imm(<4 x i32> inreg %rsrc, i32 %offset.v, i32 inreg %offset.s) {
   ; GFX6-LABEL: name: s_buffer_load_f32_offset_add_sgpr_vgpr_imm
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX6-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
   ; GFX6-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
   ; GFX6-NEXT:   [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[COPY6]], [[COPY4]], 0, implicit $exec
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GFX7-LABEL: name: s_buffer_load_f32_offset_add_sgpr_vgpr_imm
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX7-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
   ; GFX7-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
   ; GFX7-NEXT:   [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[COPY6]], [[COPY4]], 0, implicit $exec
   ; GFX7-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GFX8-LABEL: name: s_buffer_load_f32_offset_add_sgpr_vgpr_imm
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
   ; GFX8-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
   ; GFX8-NEXT:   [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[COPY6]], [[COPY4]], 0, implicit $exec
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   %offset.base = add i32 %offset.s, %offset.v
   %offset = add i32 %offset.base, 1024
   %val = call float @llvm.amdgcn.s.buffer.load.f32(<4 x i32> %rsrc, i32 %offset, i32 0)
   ret float %val
 }
 
 ; TODO: Ideally this would be reassociated to fold.
 define amdgpu_ps float @s_buffer_load_f32_offset_add_imm_sgpr_vgpr(<4 x i32> inreg %rsrc, i32 %offset.v, i32 inreg %offset.s) {
   ; GFX6-LABEL: name: s_buffer_load_f32_offset_add_imm_sgpr_vgpr
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX6-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1024
   ; GFX6-NEXT:   [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY5]], [[S_MOV_B32_]], implicit-def $scc
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GFX7-LABEL: name: s_buffer_load_f32_offset_add_imm_sgpr_vgpr
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX7-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
   ; GFX7-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1024
   ; GFX7-NEXT:   [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY5]], [[S_MOV_B32_]], implicit-def $scc
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GFX8-LABEL: name: s_buffer_load_f32_offset_add_imm_sgpr_vgpr
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1024
   ; GFX8-NEXT:   [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY5]], [[S_MOV_B32_]], implicit-def $scc
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   %offset.base = add i32 %offset.s, 1024
   %offset = add i32 %offset.base, %offset.v
   %val = call float @llvm.amdgcn.s.buffer.load.f32(<4 x i32> %rsrc, i32 %offset, i32 0)
   ret float %val
 }
 
 define amdgpu_ps float @s_buffer_load_f32_offset_add_imm_vgpr_sgpr(<4 x i32> inreg %rsrc, i32 %offset.v, i32 inreg %offset.s) {
   ; GFX6-LABEL: name: s_buffer_load_f32_offset_add_imm_vgpr_sgpr
   ; GFX6: bb.1 (%ir-block.0):
   ; GFX6-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; GFX6-NEXT: {{  $}}
   ; GFX6-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX6-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX6-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX6-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX6-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX6-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX6-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
   ; GFX6-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1024
   ; GFX6-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
   ; GFX6-NEXT:   [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[COPY4]], [[COPY6]], 0, implicit $exec
-  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX6-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX6-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX6-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GFX7-LABEL: name: s_buffer_load_f32_offset_add_imm_vgpr_sgpr
   ; GFX7: bb.1 (%ir-block.0):
   ; GFX7-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; GFX7-NEXT: {{  $}}
   ; GFX7-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX7-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX7-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX7-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX7-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX7-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX7-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
   ; GFX7-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1024
   ; GFX7-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
   ; GFX7-NEXT:   [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[COPY4]], [[COPY6]], 0, implicit $exec
-  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX7-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX7-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX7-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GFX8-LABEL: name: s_buffer_load_f32_offset_add_imm_vgpr_sgpr
   ; GFX8: bb.1 (%ir-block.0):
   ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; GFX8-NEXT: {{  $}}
   ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
   ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
   ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
   ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
   ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
   ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
   ; GFX8-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1024
   ; GFX8-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
   ; GFX8-NEXT:   [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[COPY4]], [[COPY6]], 0, implicit $exec
-  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
+  ; GFX8-NEXT:   [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32), addrspace 7)
   ; GFX8-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   %offset.base = add i32 %offset.v, 1024
   %offset = add i32 %offset.base, %offset.s
   %val = call float @llvm.amdgcn.s.buffer.load.f32(<4 x i32> %rsrc, i32 %offset, i32 0)
   ret float %val
 }
 

llvm/test/CodeGen/AMDGPU/GlobalISel/regbankselect-amdgcn-s-buffer-load.mir

bb.0:
     ; CHECK-LABEL: name: buffer_load_ss
     ; CHECK: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4
     ; CHECK-NEXT: {{  $}}
     ; CHECK-NEXT: [[COPY:%[0-9]+]]:sgpr(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
     ; CHECK-NEXT: [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
     ; CHECK-NEXT: [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:sgpr(<4 x s32>) = G_AMDGPU_S_BUFFER_LOAD [[COPY]](<4 x s32>), [[COPY1]](s32), 0
     %0:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
     %1:_(s32) = COPY $sgpr4
-    %2:_(<4 x s32>) = G_AMDGPU_S_BUFFER_LOAD %0, %1, 0
+    %2:_(<4 x s32>) = G_AMDGPU_S_BUFFER_LOAD %0, %1, 0 :: (dereferenceable invariant load (s128), align 16)
 
 ...
 
 ---
 name: buffer_load_sv
 legalized: true
 tracksRegLiveness: true
 body: |
   bb.0:
     liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $vgpr0
 
     ; CHECK-LABEL: name: buffer_load_sv
     ; CHECK: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $vgpr0
     ; CHECK-NEXT: {{  $}}
     ; CHECK-NEXT: [[COPY:%[0-9]+]]:sgpr(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
     ; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
     ; CHECK-NEXT: [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
     ; CHECK-NEXT: [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
     ; CHECK-NEXT: [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[COPY]](<4 x s32>), [[C1]](s32), [[COPY1]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
     %0:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
     %1:_(s32) = COPY $vgpr0
-    %2:_(<4 x s32>) = G_AMDGPU_S_BUFFER_LOAD %0, %1, 0
+    %2:_(<4 x s32>) = G_AMDGPU_S_BUFFER_LOAD %0, %1, 0 :: (dereferenceable invariant load (s128), align 16)
 
 ...
 
 ---
 name: buffer_load_vs
 legalized: true
 tracksRegLiveness: true
 body: |
bb.0:
     ; CHECK-NEXT: .3:
     ; CHECK-NEXT: successors: %bb.4(0x80000000)
     ; CHECK-NEXT: {{  $}}
     ; CHECK-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
     ; CHECK-NEXT: {{  $}}
     ; CHECK-NEXT: .4:
     %0:_(<4 x s32>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3
     %1:_(s32) = COPY $sgpr0
-    %2:_(<4 x s32>) = G_AMDGPU_S_BUFFER_LOAD %0, %1, 0
+    %2:_(<4 x s32>) = G_AMDGPU_S_BUFFER_LOAD %0, %1, 0 :: (dereferenceable invariant load (s128), align 16)
 
 ...
 
 ---
 name: buffer_load_vv
 legalized: true
 tracksRegLiveness: true
 body: |
bb.0:
     ; CHECK-NEXT: .3:
     ; CHECK-NEXT: successors: %bb.4(0x80000000)
     ; CHECK-NEXT: {{  $}}
     ; CHECK-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
     ; CHECK-NEXT: {{  $}}
     ; CHECK-NEXT: .4:
     %0:_(<4 x s32>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3
     %1:_(s32) = COPY $vgpr4
-    %2:_(<4 x s32>) = G_AMDGPU_S_BUFFER_LOAD %0, %1, 0
+    %2:_(<4 x s32>) = G_AMDGPU_S_BUFFER_LOAD %0, %1, 0 :: (dereferenceable invariant load (s128), align 16)
 
 ...

llvm/test/CodeGen/AMDGPU/GlobalISel/regbankselect-amdgcn.s.buffer.load.ll

 ; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 ; RUN: llc -amdgpu-global-isel-new-legality -global-isel -mtriple=amdgcn-mesa-mesa3d -mcpu=hawaii -simplify-mir -stop-after=amdgpu-regbankselect -regbankselect-fast -o - %s | FileCheck %s
 ; RUN: llc -amdgpu-global-isel-new-legality -global-isel -mtriple=amdgcn-mesa-mesa3d -mcpu=hawaii -simplify-mir -stop-after=amdgpu-regbankselect -regbankselect-greedy -o - %s | FileCheck %s --check-prefix=GREEDY
 
 ; Natural mapping
 define amdgpu_ps i32 @s_buffer_load_i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
   ; CHECK-LABEL: name: s_buffer_load_i32
   ; CHECK: bb.1 (%ir-block.0):
   ; CHECK-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:sgpr(s32) = COPY $sgpr6
-  ; CHECK-NEXT:   [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:sgpr(s32) = G_AMDGPU_S_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[COPY4]](s32), 0 :: (dereferenceable invariant load (s32))
+  ; CHECK-NEXT:   [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:sgpr(s32) = G_AMDGPU_S_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[COPY4]](s32), 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[AMDGPU_S_BUFFER_LOAD]](s32)
   ; CHECK-NEXT:   [[INT:%[0-9]+]]:sgpr(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.readfirstlane), [[COPY5]](s32)
   ; CHECK-NEXT:   $sgpr0 = COPY [[INT]](s32)
   ; CHECK-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   ; GREEDY-LABEL: name: s_buffer_load_i32
   ; GREEDY: bb.1 (%ir-block.0):
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:sgpr(s32) = COPY $sgpr6
-  ; GREEDY-NEXT:   [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:sgpr(s32) = G_AMDGPU_S_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[COPY4]](s32), 0 :: (dereferenceable invariant load (s32))
+  ; GREEDY-NEXT:   [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:sgpr(s32) = G_AMDGPU_S_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[COPY4]](s32), 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GREEDY-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[AMDGPU_S_BUFFER_LOAD]](s32)
   ; GREEDY-NEXT:   [[INT:%[0-9]+]]:sgpr(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.readfirstlane), [[COPY5]](s32)
   ; GREEDY-NEXT:   $sgpr0 = COPY [[INT]](s32)
   ; GREEDY-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0
   %val = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %rsrc, i32 %soffset, i32 0)
   ret i32 %val
 }
 
 define amdgpu_ps <2 x i32> @s_buffer_load_v2i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
   ; CHECK-LABEL: name: s_buffer_load_v2i32
   ; CHECK: bb.1 (%ir-block.0):
   ; CHECK-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:sgpr(s32) = COPY $sgpr6
-  ; CHECK-NEXT:   [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:sgpr(<2 x s32>) = G_AMDGPU_S_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[COPY4]](s32), 0 :: (dereferenceable invariant load (s64), align 4)
+  ; CHECK-NEXT:   [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:sgpr(<2 x s32>) = G_AMDGPU_S_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[COPY4]](s32), 0 :: (dereferenceable invariant load (<2 x s32>), align 4, addrspace 7)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:sgpr(s32), [[UV1:%[0-9]+]]:sgpr(s32) = G_UNMERGE_VALUES [[AMDGPU_S_BUFFER_LOAD]](<2 x s32>)
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[UV]](s32)
   ; CHECK-NEXT:   [[INT:%[0-9]+]]:sgpr(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.readfirstlane), [[COPY5]](s32)
   ; CHECK-NEXT:   $sgpr0 = COPY [[INT]](s32)
   ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:vgpr(s32) = COPY [[UV1]](s32)
   ; CHECK-NEXT:   [[INT1:%[0-9]+]]:sgpr(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.readfirstlane), [[COPY6]](s32)
   ; CHECK-NEXT:   $sgpr1 = COPY [[INT1]](s32)
   ; CHECK-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1
   ; GREEDY-LABEL: name: s_buffer_load_v2i32
   ; GREEDY: bb.1 (%ir-block.0):
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:sgpr(s32) = COPY $sgpr6
-  ; GREEDY-NEXT:   [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:sgpr(<2 x s32>) = G_AMDGPU_S_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[COPY4]](s32), 0 :: (dereferenceable invariant load (s64), align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:sgpr(<2 x s32>) = G_AMDGPU_S_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[COPY4]](s32), 0 :: (dereferenceable invariant load (<2 x s32>), align 4, addrspace 7)
   ; GREEDY-NEXT:   [[UV:%[0-9]+]]:sgpr(s32), [[UV1:%[0-9]+]]:sgpr(s32) = G_UNMERGE_VALUES [[AMDGPU_S_BUFFER_LOAD]](<2 x s32>)
   ; GREEDY-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[UV]](s32)
   ; GREEDY-NEXT:   [[INT:%[0-9]+]]:sgpr(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.readfirstlane), [[COPY5]](s32)
   ; GREEDY-NEXT:   $sgpr0 = COPY [[INT]](s32)
   ; GREEDY-NEXT:   [[COPY6:%[0-9]+]]:vgpr(s32) = COPY [[UV1]](s32)
   ; GREEDY-NEXT:   [[INT1:%[0-9]+]]:sgpr(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.readfirstlane), [[COPY6]](s32)
   ; GREEDY-NEXT:   $sgpr1 = COPY [[INT1]](s32)
   ; GREEDY-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1
   %val = call <2 x i32> @llvm.amdgcn.s.buffer.load.v2i32(<4 x i32> %rsrc, i32 %soffset, i32 0)
   ret <2 x i32> %val
 }
 
 define amdgpu_ps <3 x i32> @s_buffer_load_v3i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
   ; CHECK-LABEL: name: s_buffer_load_v3i32
   ; CHECK: bb.1 (%ir-block.0):
   ; CHECK-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:sgpr(s32) = COPY $sgpr6
-  ; CHECK-NEXT:   [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:sgpr(<4 x s32>) = G_AMDGPU_S_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[COPY4]](s32), 0 :: (dereferenceable invariant load (s96), align 4)
+  ; CHECK-NEXT:   [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:sgpr(<4 x s32>) = G_AMDGPU_S_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[COPY4]](s32), 0 :: (dereferenceable invariant load (<3 x s32>), align 4, addrspace 7)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:sgpr(s32), [[UV1:%[0-9]+]]:sgpr(s32), [[UV2:%[0-9]+]]:sgpr(s32), [[UV3:%[0-9]+]]:sgpr(s32) = G_UNMERGE_VALUES [[AMDGPU_S_BUFFER_LOAD]](<4 x s32>)
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[UV]](s32)
   ; CHECK-NEXT:   [[INT:%[0-9]+]]:sgpr(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.readfirstlane), [[COPY5]](s32)
   ; CHECK-NEXT:   $sgpr0 = COPY [[INT]](s32)
   ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:vgpr(s32) = COPY [[UV1]](s32)
   ; CHECK-NEXT:   [[INT1:%[0-9]+]]:sgpr(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.readfirstlane), [[COPY6]](s32)
   ; CHECK-NEXT:   $sgpr1 = COPY [[INT1]](s32)
   ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:vgpr(s32) = COPY [[UV2]](s32)
   ; CHECK-NEXT:   [[INT2:%[0-9]+]]:sgpr(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.readfirstlane), [[COPY7]](s32)
   ; CHECK-NEXT:   $sgpr2 = COPY [[INT2]](s32)
   ; CHECK-NEXT:   SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1, implicit $sgpr2
   ; GREEDY-LABEL: name: s_buffer_load_v3i32
   ; GREEDY: bb.1 (%ir-block.0):
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:sgpr(s32) = COPY $sgpr6
-  ; GREEDY-NEXT:   [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:sgpr(<4 x s32>) = G_AMDGPU_S_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[COPY4]](s32), 0 :: (dereferenceable invariant load (s96), align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:sgpr(<4 x s32>) = G_AMDGPU_S_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[COPY4]](s32), 0 :: (dereferenceable invariant load (<3 x s32>), align 4, addrspace 7)
   ; GREEDY-NEXT:   [[UV:%[0-9]+]]:sgpr(s32), [[UV1:%[0-9]+]]:sgpr(s32), [[UV2:%[0-9]+]]:sgpr(s32), [[UV3:%[0-9]+]]:sgpr(s32) = G_UNMERGE_VALUES [[AMDGPU_S_BUFFER_LOAD]](<4 x s32>)
   ; GREEDY-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[UV]](s32)
   ; GREEDY-NEXT:   [[INT:%[0-9]+]]:sgpr(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.readfirstlane), [[COPY5]](s32)
   ; GREEDY-NEXT:   $sgpr0 = COPY [[INT]](s32)
   ; GREEDY-NEXT:   [[COPY6:%[0-9]+]]:vgpr(s32) = COPY [[UV1]](s32)
   ; GREEDY-NEXT:   [[INT1:%[0-9]+]]:sgpr(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.readfirstlane), [[COPY6]](s32)
   ; GREEDY-NEXT:   $sgpr1 = COPY [[INT1]](s32)
   ; GREEDY-NEXT:   [[COPY7:%[0-9]+]]:vgpr(s32) = COPY [[UV2]](s32)
define amdgpu_ps <8 x i32> @s_buffer_load_v8i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
   ; CHECK-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:sgpr(s32) = COPY $sgpr6
-  ; CHECK-NEXT:   [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:sgpr(<8 x s32>) = G_AMDGPU_S_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[COPY4]](s32), 0 :: (dereferenceable invariant load (s256), align 4)
+  ; CHECK-NEXT:   [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:sgpr(<8 x s32>) = G_AMDGPU_S_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[COPY4]](s32), 0 :: (dereferenceable invariant load (<8 x s32>), align 4, addrspace 7)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:sgpr(s32), [[UV1:%[0-9]+]]:sgpr(s32), [[UV2:%[0-9]+]]:sgpr(s32), [[UV3:%[0-9]+]]:sgpr(s32), [[UV4:%[0-9]+]]:sgpr(s32), [[UV5:%[0-9]+]]:sgpr(s32), [[UV6:%[0-9]+]]:sgpr(s32), [[UV7:%[0-9]+]]:sgpr(s32) = G_UNMERGE_VALUES [[AMDGPU_S_BUFFER_LOAD]](<8 x s32>)
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[UV]](s32)
   ; CHECK-NEXT:   [[INT:%[0-9]+]]:sgpr(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.readfirstlane), [[COPY5]](s32)
   ; CHECK-NEXT:   $sgpr0 = COPY [[INT]](s32)
   ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:vgpr(s32) = COPY [[UV1]](s32)
   ; CHECK-NEXT:   [[INT1:%[0-9]+]]:sgpr(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.readfirstlane), [[COPY6]](s32)
   ; CHECK-NEXT:   $sgpr1 = COPY [[INT1]](s32)
   ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:vgpr(s32) = COPY [[UV2]](s32)
define amdgpu_ps <8 x i32> @s_buffer_load_v8i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:sgpr(s32) = COPY $sgpr6
-  ; GREEDY-NEXT:   [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:sgpr(<8 x s32>) = G_AMDGPU_S_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[COPY4]](s32), 0 :: (dereferenceable invariant load (s256), align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:sgpr(<8 x s32>) = G_AMDGPU_S_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[COPY4]](s32), 0 :: (dereferenceable invariant load (<8 x s32>), align 4, addrspace 7)
   ; GREEDY-NEXT:   [[UV:%[0-9]+]]:sgpr(s32), [[UV1:%[0-9]+]]:sgpr(s32), [[UV2:%[0-9]+]]:sgpr(s32), [[UV3:%[0-9]+]]:sgpr(s32), [[UV4:%[0-9]+]]:sgpr(s32), [[UV5:%[0-9]+]]:sgpr(s32), [[UV6:%[0-9]+]]:sgpr(s32), [[UV7:%[0-9]+]]:sgpr(s32) = G_UNMERGE_VALUES [[AMDGPU_S_BUFFER_LOAD]](<8 x s32>)
   ; GREEDY-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[UV]](s32)
   ; GREEDY-NEXT:   [[INT:%[0-9]+]]:sgpr(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.readfirstlane), [[COPY5]](s32)
   ; GREEDY-NEXT:   $sgpr0 = COPY [[INT]](s32)
   ; GREEDY-NEXT:   [[COPY6:%[0-9]+]]:vgpr(s32) = COPY [[UV1]](s32)
   ; GREEDY-NEXT:   [[INT1:%[0-9]+]]:sgpr(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.readfirstlane), [[COPY6]](s32)
   ; GREEDY-NEXT:   $sgpr1 = COPY [[INT1]](s32)
   ; GREEDY-NEXT:   [[COPY7:%[0-9]+]]:vgpr(s32) = COPY [[UV2]](s32)
define amdgpu_ps <16 x i32> @s_buffer_load_v16i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
   ; CHECK-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:sgpr(s32) = COPY $sgpr6
-  ; CHECK-NEXT:   [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:sgpr(<16 x s32>) = G_AMDGPU_S_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[COPY4]](s32), 0 :: (dereferenceable invariant load (s512), align 4)
+  ; CHECK-NEXT:   [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:sgpr(<16 x s32>) = G_AMDGPU_S_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[COPY4]](s32), 0 :: (dereferenceable invariant load (<16 x s32>), align 4, addrspace 7)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:sgpr(s32), [[UV1:%[0-9]+]]:sgpr(s32), [[UV2:%[0-9]+]]:sgpr(s32), [[UV3:%[0-9]+]]:sgpr(s32), [[UV4:%[0-9]+]]:sgpr(s32), [[UV5:%[0-9]+]]:sgpr(s32), [[UV6:%[0-9]+]]:sgpr(s32), [[UV7:%[0-9]+]]:sgpr(s32), [[UV8:%[0-9]+]]:sgpr(s32), [[UV9:%[0-9]+]]:sgpr(s32), [[UV10:%[0-9]+]]:sgpr(s32), [[UV11:%[0-9]+]]:sgpr(s32), [[UV12:%[0-9]+]]:sgpr(s32), [[UV13:%[0-9]+]]:sgpr(s32), [[UV14:%[0-9]+]]:sgpr(s32), [[UV15:%[0-9]+]]:sgpr(s32) = G_UNMERGE_VALUES [[AMDGPU_S_BUFFER_LOAD]](<16 x s32>)
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[UV]](s32)
   ; CHECK-NEXT:   [[INT:%[0-9]+]]:sgpr(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.readfirstlane), [[COPY5]](s32)
   ; CHECK-NEXT:   $sgpr0 = COPY [[INT]](s32)
   ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:vgpr(s32) = COPY [[UV1]](s32)
   ; CHECK-NEXT:   [[INT1:%[0-9]+]]:sgpr(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.readfirstlane), [[COPY6]](s32)
   ; CHECK-NEXT:   $sgpr1 = COPY [[INT1]](s32)
   ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:vgpr(s32) = COPY [[UV2]](s32)
define amdgpu_ps <16 x i32> @s_buffer_load_v16i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:sgpr(s32) = COPY $sgpr6
-  ; GREEDY-NEXT:   [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:sgpr(<16 x s32>) = G_AMDGPU_S_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[COPY4]](s32), 0 :: (dereferenceable invariant load (s512), align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_S_BUFFER_LOAD:%[0-9]+]]:sgpr(<16 x s32>) = G_AMDGPU_S_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[COPY4]](s32), 0 :: (dereferenceable invariant load (<16 x s32>), align 4, addrspace 7)
   ; GREEDY-NEXT:   [[UV:%[0-9]+]]:sgpr(s32), [[UV1:%[0-9]+]]:sgpr(s32), [[UV2:%[0-9]+]]:sgpr(s32), [[UV3:%[0-9]+]]:sgpr(s32), [[UV4:%[0-9]+]]:sgpr(s32), [[UV5:%[0-9]+]]:sgpr(s32), [[UV6:%[0-9]+]]:sgpr(s32), [[UV7:%[0-9]+]]:sgpr(s32), [[UV8:%[0-9]+]]:sgpr(s32), [[UV9:%[0-9]+]]:sgpr(s32), [[UV10:%[0-9]+]]:sgpr(s32), [[UV11:%[0-9]+]]:sgpr(s32), [[UV12:%[0-9]+]]:sgpr(s32), [[UV13:%[0-9]+]]:sgpr(s32), [[UV14:%[0-9]+]]:sgpr(s32), [[UV15:%[0-9]+]]:sgpr(s32) = G_UNMERGE_VALUES [[AMDGPU_S_BUFFER_LOAD]](<16 x s32>)
   ; GREEDY-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[UV]](s32)
   ; GREEDY-NEXT:   [[INT:%[0-9]+]]:sgpr(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.readfirstlane), [[COPY5]](s32)
   ; GREEDY-NEXT:   $sgpr0 = COPY [[INT]](s32)
   ; GREEDY-NEXT:   [[COPY6:%[0-9]+]]:vgpr(s32) = COPY [[UV1]](s32)
   ; GREEDY-NEXT:   [[INT1:%[0-9]+]]:sgpr(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.readfirstlane), [[COPY6]](s32)
   ; GREEDY-NEXT:   $sgpr1 = COPY [[INT1]](s32)
   ; GREEDY-NEXT:   [[COPY7:%[0-9]+]]:vgpr(s32) = COPY [[UV2]](s32)
define amdgpu_ps float @s_buffer_load_f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s32))
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; CHECK-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
   ; CHECK-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GREEDY-LABEL: name: s_buffer_load_f32_vgpr_offset
   ; GREEDY: bb.1 (%ir-block.0):
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s32))
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GREEDY-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
   ; GREEDY-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   %val = call float @llvm.amdgcn.s.buffer.load.f32(<4 x i32> %rsrc, i32 %soffset, i32 0)
   ret float %val
 }
 
 define amdgpu_ps <2 x float> @s_buffer_load_v2f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; CHECK-LABEL: name: s_buffer_load_v2f32_vgpr_offset
   ; CHECK: bb.1 (%ir-block.0):
   ; CHECK-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<2 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s64), align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<2 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s64), align 4, addrspace 7)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:vgpr(s32), [[UV1:%[0-9]+]]:vgpr(s32) = G_UNMERGE_VALUES [[AMDGPU_BUFFER_LOAD]](<2 x s32>)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; CHECK-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1
   ; GREEDY-LABEL: name: s_buffer_load_v2f32_vgpr_offset
   ; GREEDY: bb.1 (%ir-block.0):
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<2 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s64), align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<2 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s64), align 4, addrspace 7)
   ; GREEDY-NEXT:   [[UV:%[0-9]+]]:vgpr(s32), [[UV1:%[0-9]+]]:vgpr(s32) = G_UNMERGE_VALUES [[AMDGPU_BUFFER_LOAD]](<2 x s32>)
   ; GREEDY-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; GREEDY-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; GREEDY-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1
   %val = call <2 x float> @llvm.amdgcn.s.buffer.load.v2f32(<4 x i32> %rsrc, i32 %soffset, i32 0)
   ret <2 x float> %val
 }
 
 define amdgpu_ps <3 x float> @s_buffer_load_v3f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; CHECK-LABEL: name: s_buffer_load_v3f32_vgpr_offset
   ; CHECK: bb.1 (%ir-block.0):
   ; CHECK-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:vgpr(s32), [[UV1:%[0-9]+]]:vgpr(s32), [[UV2:%[0-9]+]]:vgpr(s32), [[UV3:%[0-9]+]]:vgpr(s32) = G_UNMERGE_VALUES [[AMDGPU_BUFFER_LOAD]](<4 x s32>)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; CHECK-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
   ; GREEDY-LABEL: name: s_buffer_load_v3f32_vgpr_offset
   ; GREEDY: bb.1 (%ir-block.0):
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GREEDY-NEXT:   [[UV:%[0-9]+]]:vgpr(s32), [[UV1:%[0-9]+]]:vgpr(s32), [[UV2:%[0-9]+]]:vgpr(s32), [[UV3:%[0-9]+]]:vgpr(s32) = G_UNMERGE_VALUES [[AMDGPU_BUFFER_LOAD]](<4 x s32>)
   ; GREEDY-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; GREEDY-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; GREEDY-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; GREEDY-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
   %val = call <3 x float> @llvm.amdgcn.s.buffer.load.v3f32(<4 x i32> %rsrc, i32 %soffset, i32 0)
   ret <3 x float> %val
 }
 
 define amdgpu_ps <4 x float> @s_buffer_load_v4f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; CHECK-LABEL: name: s_buffer_load_v4f32_vgpr_offset
   ; CHECK: bb.1 (%ir-block.0):
   ; CHECK-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:vgpr(s32), [[UV1:%[0-9]+]]:vgpr(s32), [[UV2:%[0-9]+]]:vgpr(s32), [[UV3:%[0-9]+]]:vgpr(s32) = G_UNMERGE_VALUES [[AMDGPU_BUFFER_LOAD]](<4 x s32>)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
   ; CHECK-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
   ; GREEDY-LABEL: name: s_buffer_load_v4f32_vgpr_offset
   ; GREEDY: bb.1 (%ir-block.0):
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GREEDY-NEXT:   [[UV:%[0-9]+]]:vgpr(s32), [[UV1:%[0-9]+]]:vgpr(s32), [[UV2:%[0-9]+]]:vgpr(s32), [[UV3:%[0-9]+]]:vgpr(s32) = G_UNMERGE_VALUES [[AMDGPU_BUFFER_LOAD]](<4 x s32>)
   ; GREEDY-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; GREEDY-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; GREEDY-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; GREEDY-NEXT:   $vgpr3 = COPY [[UV3]](s32)
   ; GREEDY-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
   %val = call <4 x float> @llvm.amdgcn.s.buffer.load.v4f32(<4 x i32> %rsrc, i32 %soffset, i32 0)
   ret <4 x float> %val
 }
 
 define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; CHECK-LABEL: name: s_buffer_load_v8f32_vgpr_offset
   ; CHECK: bb.1 (%ir-block.0):
   ; CHECK-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; CHECK-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:vgpr(s32), [[UV1:%[0-9]+]]:vgpr(s32), [[UV2:%[0-9]+]]:vgpr(s32), [[UV3:%[0-9]+]]:vgpr(s32), [[UV4:%[0-9]+]]:vgpr(s32), [[UV5:%[0-9]+]]:vgpr(s32), [[UV6:%[0-9]+]]:vgpr(s32), [[UV7:%[0-9]+]]:vgpr(s32) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<8 x s32>)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
   ; CHECK-NEXT:   $vgpr4 = COPY [[UV4]](s32)
   ; CHECK-NEXT:   $vgpr5 = COPY [[UV5]](s32)
   ; CHECK-NEXT:   $vgpr6 = COPY [[UV6]](s32)
   ; CHECK-NEXT:   $vgpr7 = COPY [[UV7]](s32)
   ; CHECK-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
   ; GREEDY-LABEL: name: s_buffer_load_v8f32_vgpr_offset
   ; GREEDY: bb.1 (%ir-block.0):
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GREEDY-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>)
   ; GREEDY-NEXT:   [[UV:%[0-9]+]]:vgpr(s32), [[UV1:%[0-9]+]]:vgpr(s32), [[UV2:%[0-9]+]]:vgpr(s32), [[UV3:%[0-9]+]]:vgpr(s32), [[UV4:%[0-9]+]]:vgpr(s32), [[UV5:%[0-9]+]]:vgpr(s32), [[UV6:%[0-9]+]]:vgpr(s32), [[UV7:%[0-9]+]]:vgpr(s32) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<8 x s32>)
   ; GREEDY-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; GREEDY-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; GREEDY-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; GREEDY-NEXT:   $vgpr3 = COPY [[UV3]](s32)
   ; GREEDY-NEXT:   $vgpr4 = COPY [[UV4]](s32)
   ; GREEDY-NEXT:   $vgpr5 = COPY [[UV5]](s32)
define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4, addrspace 7)
   ; CHECK-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<16 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>), [[AMDGPU_BUFFER_LOAD2]](<4 x s32>), [[AMDGPU_BUFFER_LOAD3]](<4 x s32>)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:vgpr(s32), [[UV1:%[0-9]+]]:vgpr(s32), [[UV2:%[0-9]+]]:vgpr(s32), [[UV3:%[0-9]+]]:vgpr(s32), [[UV4:%[0-9]+]]:vgpr(s32), [[UV5:%[0-9]+]]:vgpr(s32), [[UV6:%[0-9]+]]:vgpr(s32), [[UV7:%[0-9]+]]:vgpr(s32), [[UV8:%[0-9]+]]:vgpr(s32), [[UV9:%[0-9]+]]:vgpr(s32), [[UV10:%[0-9]+]]:vgpr(s32), [[UV11:%[0-9]+]]:vgpr(s32), [[UV12:%[0-9]+]]:vgpr(s32), [[UV13:%[0-9]+]]:vgpr(s32), [[UV14:%[0-9]+]]:vgpr(s32), [[UV15:%[0-9]+]]:vgpr(s32) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<16 x s32>)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
   ; CHECK-NEXT:   $vgpr4 = COPY [[UV4]](s32)
   ; CHECK-NEXT:   $vgpr5 = COPY [[UV5]](s32)
define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4, addrspace 7)
   ; GREEDY-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<16 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>), [[AMDGPU_BUFFER_LOAD2]](<4 x s32>), [[AMDGPU_BUFFER_LOAD3]](<4 x s32>)
   ; GREEDY-NEXT:   [[UV:%[0-9]+]]:vgpr(s32), [[UV1:%[0-9]+]]:vgpr(s32), [[UV2:%[0-9]+]]:vgpr(s32), [[UV3:%[0-9]+]]:vgpr(s32), [[UV4:%[0-9]+]]:vgpr(s32), [[UV5:%[0-9]+]]:vgpr(s32), [[UV6:%[0-9]+]]:vgpr(s32), [[UV7:%[0-9]+]]:vgpr(s32), [[UV8:%[0-9]+]]:vgpr(s32), [[UV9:%[0-9]+]]:vgpr(s32), [[UV10:%[0-9]+]]:vgpr(s32), [[UV11:%[0-9]+]]:vgpr(s32), [[UV12:%[0-9]+]]:vgpr(s32), [[UV13:%[0-9]+]]:vgpr(s32), [[UV14:%[0-9]+]]:vgpr(s32), [[UV15:%[0-9]+]]:vgpr(s32) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<16 x s32>)
   ; GREEDY-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; GREEDY-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; GREEDY-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; GREEDY-NEXT:   $vgpr3 = COPY [[UV3]](s32)
   ; GREEDY-NEXT:   $vgpr4 = COPY [[UV4]](s32)
   ; GREEDY-NEXT:   $vgpr5 = COPY [[UV5]](s32)
define amdgpu_ps void @s_buffer_load_i96_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[DEF:%[0-9]+]]:sgpr(p1) = G_IMPLICIT_DEF
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; CHECK-NEXT:   [[TRUNC:%[0-9]+]]:vgpr(s96) = G_TRUNC [[AMDGPU_BUFFER_LOAD]](s128)
   ; CHECK-NEXT:   G_STORE [[TRUNC]](s96), [[DEF]](p1) :: (store (s96) into `ptr addrspace(1) undef`, align 8, addrspace 1)
   ; CHECK-NEXT:   S_ENDPGM 0
   ; GREEDY-LABEL: name: s_buffer_load_i96_vgpr_offset
   ; GREEDY: bb.1 (%ir-block.0):
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[DEF:%[0-9]+]]:sgpr(p1) = G_IMPLICIT_DEF
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GREEDY-NEXT:   [[TRUNC:%[0-9]+]]:vgpr(s96) = G_TRUNC [[AMDGPU_BUFFER_LOAD]](s128)
   ; GREEDY-NEXT:   G_STORE [[TRUNC]](s96), [[DEF]](p1) :: (store (s96) into `ptr addrspace(1) undef`, align 8, addrspace 1)
   ; GREEDY-NEXT:   S_ENDPGM 0
   %val = call i96 @llvm.amdgcn.s.buffer.load.i96(<4 x i32> %rsrc, i32 %soffset, i32 0)
   store i96 %val, ptr addrspace(1) undef
   ret void
 }
 
 ; Test split of a wide scalar
 define amdgpu_ps void @s_buffer_load_i256_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; CHECK-LABEL: name: s_buffer_load_i256_vgpr_offset
   ; CHECK: bb.1 (%ir-block.0):
   ; CHECK-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[DEF:%[0-9]+]]:sgpr(p1) = G_IMPLICIT_DEF
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; CHECK-NEXT:   [[MV:%[0-9]+]]:vgpr(s256) = G_MERGE_VALUES [[AMDGPU_BUFFER_LOAD]](s128), [[AMDGPU_BUFFER_LOAD1]](s128)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:vgpr(s128), [[UV1:%[0-9]+]]:vgpr(s128) = G_UNMERGE_VALUES [[MV]](s256)
   ; CHECK-NEXT:   G_STORE [[UV]](s128), [[DEF]](p1) :: (store (s128) into `ptr addrspace(1) undef`, align 8, addrspace 1)
   ; CHECK-NEXT:   [[C2:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 16
   ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C2]](s64)
   ; CHECK-NEXT:   G_STORE [[UV1]](s128), [[PTR_ADD]](p1) :: (store (s128) into `ptr addrspace(1) undef` + 16, align 8, addrspace 1)
   ; CHECK-NEXT:   S_ENDPGM 0
   ; GREEDY-LABEL: name: s_buffer_load_i256_vgpr_offset
   ; GREEDY: bb.1 (%ir-block.0):
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[DEF:%[0-9]+]]:sgpr(p1) = G_IMPLICIT_DEF
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GREEDY-NEXT:   [[MV:%[0-9]+]]:vgpr(s256) = G_MERGE_VALUES [[AMDGPU_BUFFER_LOAD]](s128), [[AMDGPU_BUFFER_LOAD1]](s128)
   ; GREEDY-NEXT:   [[UV:%[0-9]+]]:vgpr(s128), [[UV1:%[0-9]+]]:vgpr(s128) = G_UNMERGE_VALUES [[MV]](s256)
   ; GREEDY-NEXT:   G_STORE [[UV]](s128), [[DEF]](p1) :: (store (s128) into `ptr addrspace(1) undef`, align 8, addrspace 1)
   ; GREEDY-NEXT:   [[C2:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 16
   ; GREEDY-NEXT:   [[PTR_ADD:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C2]](s64)
   ; GREEDY-NEXT:   G_STORE [[UV1]](s128), [[PTR_ADD]](p1) :: (store (s128) into `ptr addrspace(1) undef` + 16, align 8, addrspace 1)
   ; GREEDY-NEXT:   S_ENDPGM 0
   %val = call i256 @llvm.amdgcn.s.buffer.load.i256(<4 x i32> %rsrc, i32 %soffset, i32 0)
define amdgpu_ps void @s_buffer_load_i512_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[DEF:%[0-9]+]]:sgpr(p1) = G_IMPLICIT_DEF
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4, addrspace 7)
   ; CHECK-NEXT:   [[MV:%[0-9]+]]:vgpr(s512) = G_MERGE_VALUES [[AMDGPU_BUFFER_LOAD]](s128), [[AMDGPU_BUFFER_LOAD1]](s128), [[AMDGPU_BUFFER_LOAD2]](s128), [[AMDGPU_BUFFER_LOAD3]](s128)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:vgpr(s128), [[UV1:%[0-9]+]]:vgpr(s128), [[UV2:%[0-9]+]]:vgpr(s128), [[UV3:%[0-9]+]]:vgpr(s128) = G_UNMERGE_VALUES [[MV]](s512)
   ; CHECK-NEXT:   G_STORE [[UV]](s128), [[DEF]](p1) :: (store (s128) into `ptr addrspace(1) undef`, align 8, addrspace 1)
   ; CHECK-NEXT:   [[C2:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 16
   ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C2]](s64)
   ; CHECK-NEXT:   G_STORE [[UV1]](s128), [[PTR_ADD]](p1) :: (store (s128) into `ptr addrspace(1) undef` + 16, align 8, addrspace 1)
   ; CHECK-NEXT:   [[C3:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 32
   ; CHECK-NEXT:   [[PTR_ADD1:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C3]](s64)
define amdgpu_ps void @s_buffer_load_i512_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[DEF:%[0-9]+]]:sgpr(p1) = G_IMPLICIT_DEF
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4, addrspace 7)
   ; GREEDY-NEXT:   [[MV:%[0-9]+]]:vgpr(s512) = G_MERGE_VALUES [[AMDGPU_BUFFER_LOAD]](s128), [[AMDGPU_BUFFER_LOAD1]](s128), [[AMDGPU_BUFFER_LOAD2]](s128), [[AMDGPU_BUFFER_LOAD3]](s128)
   ; GREEDY-NEXT:   [[UV:%[0-9]+]]:vgpr(s128), [[UV1:%[0-9]+]]:vgpr(s128), [[UV2:%[0-9]+]]:vgpr(s128), [[UV3:%[0-9]+]]:vgpr(s128) = G_UNMERGE_VALUES [[MV]](s512)
   ; GREEDY-NEXT:   G_STORE [[UV]](s128), [[DEF]](p1) :: (store (s128) into `ptr addrspace(1) undef`, align 8, addrspace 1)
   ; GREEDY-NEXT:   [[C2:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 16
   ; GREEDY-NEXT:   [[PTR_ADD:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C2]](s64)
   ; GREEDY-NEXT:   G_STORE [[UV1]](s128), [[PTR_ADD]](p1) :: (store (s128) into `ptr addrspace(1) undef` + 16, align 8, addrspace 1)
   ; GREEDY-NEXT:   [[C3:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 32
   ; GREEDY-NEXT:   [[PTR_ADD1:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C3]](s64)
define amdgpu_ps void @s_buffer_load_v16i16_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[DEF:%[0-9]+]]:sgpr(p1) = G_IMPLICIT_DEF
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; CHECK-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<16 x s16>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<8 x s16>), [[AMDGPU_BUFFER_LOAD1]](<8 x s16>)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:vgpr(<8 x s16>), [[UV1:%[0-9]+]]:vgpr(<8 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<16 x s16>)
   ; CHECK-NEXT:   G_STORE [[UV]](<8 x s16>), [[DEF]](p1) :: (store (<8 x s16>) into `ptr addrspace(1) undef`, align 32, addrspace 1)
   ; CHECK-NEXT:   [[C2:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 16
   ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C2]](s64)
   ; CHECK-NEXT:   G_STORE [[UV1]](<8 x s16>), [[PTR_ADD]](p1) :: (store (<8 x s16>) into `ptr addrspace(1) undef` + 16, basealign 32, addrspace 1)
   ; CHECK-NEXT:   S_ENDPGM 0
   ; GREEDY-LABEL: name: s_buffer_load_v16i16_vgpr_offset
   ; GREEDY: bb.1 (%ir-block.0):
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[DEF:%[0-9]+]]:sgpr(p1) = G_IMPLICIT_DEF
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GREEDY-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<16 x s16>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<8 x s16>), [[AMDGPU_BUFFER_LOAD1]](<8 x s16>)
   ; GREEDY-NEXT:   [[UV:%[0-9]+]]:vgpr(<8 x s16>), [[UV1:%[0-9]+]]:vgpr(<8 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<16 x s16>)
   ; GREEDY-NEXT:   G_STORE [[UV]](<8 x s16>), [[DEF]](p1) :: (store (<8 x s16>) into `ptr addrspace(1) undef`, align 32, addrspace 1)
   ; GREEDY-NEXT:   [[C2:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 16
   ; GREEDY-NEXT:   [[PTR_ADD:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C2]](s64)
   ; GREEDY-NEXT:   G_STORE [[UV1]](<8 x s16>), [[PTR_ADD]](p1) :: (store (<8 x s16>) into `ptr addrspace(1) undef` + 16, basealign 32, addrspace 1)
   ; GREEDY-NEXT:   S_ENDPGM 0
   %val = call <16 x i16> @llvm.amdgcn.s.buffer.load.v16i16(<4 x i32> %rsrc, i32 %soffset, i32 0)
define amdgpu_ps void @s_buffer_load_v32i16_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[DEF:%[0-9]+]]:sgpr(p1) = G_IMPLICIT_DEF
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4, addrspace 7)
   ; CHECK-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<32 x s16>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<8 x s16>), [[AMDGPU_BUFFER_LOAD1]](<8 x s16>), [[AMDGPU_BUFFER_LOAD2]](<8 x s16>), [[AMDGPU_BUFFER_LOAD3]](<8 x s16>)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:vgpr(<8 x s16>), [[UV1:%[0-9]+]]:vgpr(<8 x s16>), [[UV2:%[0-9]+]]:vgpr(<8 x s16>), [[UV3:%[0-9]+]]:vgpr(<8 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<32 x s16>)
   ; CHECK-NEXT:   G_STORE [[UV]](<8 x s16>), [[DEF]](p1) :: (store (<8 x s16>) into `ptr addrspace(1) undef`, align 64, addrspace 1)
   ; CHECK-NEXT:   [[C2:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 16
   ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C2]](s64)
   ; CHECK-NEXT:   G_STORE [[UV1]](<8 x s16>), [[PTR_ADD]](p1) :: (store (<8 x s16>) into `ptr addrspace(1) undef` + 16, basealign 64, addrspace 1)
   ; CHECK-NEXT:   [[C3:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 32
   ; CHECK-NEXT:   [[PTR_ADD1:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C3]](s64)
define amdgpu_ps void @s_buffer_load_v32i16_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[DEF:%[0-9]+]]:sgpr(p1) = G_IMPLICIT_DEF
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4, addrspace 7)
   ; GREEDY-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<32 x s16>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<8 x s16>), [[AMDGPU_BUFFER_LOAD1]](<8 x s16>), [[AMDGPU_BUFFER_LOAD2]](<8 x s16>), [[AMDGPU_BUFFER_LOAD3]](<8 x s16>)
   ; GREEDY-NEXT:   [[UV:%[0-9]+]]:vgpr(<8 x s16>), [[UV1:%[0-9]+]]:vgpr(<8 x s16>), [[UV2:%[0-9]+]]:vgpr(<8 x s16>), [[UV3:%[0-9]+]]:vgpr(<8 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<32 x s16>)
   ; GREEDY-NEXT:   G_STORE [[UV]](<8 x s16>), [[DEF]](p1) :: (store (<8 x s16>) into `ptr addrspace(1) undef`, align 64, addrspace 1)
   ; GREEDY-NEXT:   [[C2:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 16
   ; GREEDY-NEXT:   [[PTR_ADD:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C2]](s64)
   ; GREEDY-NEXT:   G_STORE [[UV1]](<8 x s16>), [[PTR_ADD]](p1) :: (store (<8 x s16>) into `ptr addrspace(1) undef` + 16, basealign 64, addrspace 1)
   ; GREEDY-NEXT:   [[C3:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 32
   ; GREEDY-NEXT:   [[PTR_ADD1:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C3]](s64)
define amdgpu_ps void @s_buffer_load_v4i64_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[DEF:%[0-9]+]]:sgpr(p1) = G_IMPLICIT_DEF
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; CHECK-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<4 x s64>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<2 x s64>), [[AMDGPU_BUFFER_LOAD1]](<2 x s64>)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:vgpr(<2 x s64>), [[UV1:%[0-9]+]]:vgpr(<2 x s64>) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<4 x s64>)
   ; CHECK-NEXT:   G_STORE [[UV]](<2 x s64>), [[DEF]](p1) :: (store (<2 x s64>) into `ptr addrspace(1) undef`, align 32, addrspace 1)
   ; CHECK-NEXT:   [[C2:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 16
   ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C2]](s64)
   ; CHECK-NEXT:   G_STORE [[UV1]](<2 x s64>), [[PTR_ADD]](p1) :: (store (<2 x s64>) into `ptr addrspace(1) undef` + 16, basealign 32, addrspace 1)
   ; CHECK-NEXT:   S_ENDPGM 0
   ; GREEDY-LABEL: name: s_buffer_load_v4i64_vgpr_offset
   ; GREEDY: bb.1 (%ir-block.0):
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[DEF:%[0-9]+]]:sgpr(p1) = G_IMPLICIT_DEF
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GREEDY-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<4 x s64>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<2 x s64>), [[AMDGPU_BUFFER_LOAD1]](<2 x s64>)
   ; GREEDY-NEXT:   [[UV:%[0-9]+]]:vgpr(<2 x s64>), [[UV1:%[0-9]+]]:vgpr(<2 x s64>) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<4 x s64>)
   ; GREEDY-NEXT:   G_STORE [[UV]](<2 x s64>), [[DEF]](p1) :: (store (<2 x s64>) into `ptr addrspace(1) undef`, align 32, addrspace 1)
   ; GREEDY-NEXT:   [[C2:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 16
   ; GREEDY-NEXT:   [[PTR_ADD:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C2]](s64)
   ; GREEDY-NEXT:   G_STORE [[UV1]](<2 x s64>), [[PTR_ADD]](p1) :: (store (<2 x s64>) into `ptr addrspace(1) undef` + 16, basealign 32, addrspace 1)
   ; GREEDY-NEXT:   S_ENDPGM 0
   %val = call <4 x i64> @llvm.amdgcn.s.buffer.load.v4i64(<4 x i32> %rsrc, i32 %soffset, i32 0)
define amdgpu_ps void @s_buffer_load_v8i64_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[DEF:%[0-9]+]]:sgpr(p1) = G_IMPLICIT_DEF
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4, addrspace 7)
   ; CHECK-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x s64>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<2 x s64>), [[AMDGPU_BUFFER_LOAD1]](<2 x s64>), [[AMDGPU_BUFFER_LOAD2]](<2 x s64>), [[AMDGPU_BUFFER_LOAD3]](<2 x s64>)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:vgpr(<2 x s64>), [[UV1:%[0-9]+]]:vgpr(<2 x s64>), [[UV2:%[0-9]+]]:vgpr(<2 x s64>), [[UV3:%[0-9]+]]:vgpr(<2 x s64>) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<8 x s64>)
   ; CHECK-NEXT:   G_STORE [[UV]](<2 x s64>), [[DEF]](p1) :: (store (<2 x s64>) into `ptr addrspace(1) undef`, align 64, addrspace 1)
   ; CHECK-NEXT:   [[C2:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 16
   ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C2]](s64)
   ; CHECK-NEXT:   G_STORE [[UV1]](<2 x s64>), [[PTR_ADD]](p1) :: (store (<2 x s64>) into `ptr addrspace(1) undef` + 16, basealign 64, addrspace 1)
   ; CHECK-NEXT:   [[C3:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 32
   ; CHECK-NEXT:   [[PTR_ADD1:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C3]](s64)
define amdgpu_ps void @s_buffer_load_v8i64_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[DEF:%[0-9]+]]:sgpr(p1) = G_IMPLICIT_DEF
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4, addrspace 7)
   ; GREEDY-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x s64>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<2 x s64>), [[AMDGPU_BUFFER_LOAD1]](<2 x s64>), [[AMDGPU_BUFFER_LOAD2]](<2 x s64>), [[AMDGPU_BUFFER_LOAD3]](<2 x s64>)
   ; GREEDY-NEXT:   [[UV:%[0-9]+]]:vgpr(<2 x s64>), [[UV1:%[0-9]+]]:vgpr(<2 x s64>), [[UV2:%[0-9]+]]:vgpr(<2 x s64>), [[UV3:%[0-9]+]]:vgpr(<2 x s64>) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<8 x s64>)
   ; GREEDY-NEXT:   G_STORE [[UV]](<2 x s64>), [[DEF]](p1) :: (store (<2 x s64>) into `ptr addrspace(1) undef`, align 64, addrspace 1)
   ; GREEDY-NEXT:   [[C2:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 16
   ; GREEDY-NEXT:   [[PTR_ADD:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C2]](s64)
   ; GREEDY-NEXT:   G_STORE [[UV1]](<2 x s64>), [[PTR_ADD]](p1) :: (store (<2 x s64>) into `ptr addrspace(1) undef` + 16, basealign 64, addrspace 1)
   ; GREEDY-NEXT:   [[C3:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 32
   ; GREEDY-NEXT:   [[PTR_ADD1:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C3]](s64)
define amdgpu_ps void @s_buffer_load_v4p1_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[DEF:%[0-9]+]]:sgpr(p1) = G_IMPLICIT_DEF
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; CHECK-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<4 x p1>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<2 x p1>), [[AMDGPU_BUFFER_LOAD1]](<2 x p1>)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:vgpr(<2 x p1>), [[UV1:%[0-9]+]]:vgpr(<2 x p1>) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<4 x p1>)
   ; CHECK-NEXT:   G_STORE [[UV]](<2 x p1>), [[DEF]](p1) :: (store (<2 x p1>) into `ptr addrspace(1) undef`, align 32, addrspace 1)
   ; CHECK-NEXT:   [[C2:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 16
   ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C2]](s64)
   ; CHECK-NEXT:   G_STORE [[UV1]](<2 x p1>), [[PTR_ADD]](p1) :: (store (<2 x p1>) into `ptr addrspace(1) undef` + 16, basealign 32, addrspace 1)
   ; CHECK-NEXT:   S_ENDPGM 0
   ; GREEDY-LABEL: name: s_buffer_load_v4p1_vgpr_offset
   ; GREEDY: bb.1 (%ir-block.0):
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[DEF:%[0-9]+]]:sgpr(p1) = G_IMPLICIT_DEF
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GREEDY-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<4 x p1>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<2 x p1>), [[AMDGPU_BUFFER_LOAD1]](<2 x p1>)
   ; GREEDY-NEXT:   [[UV:%[0-9]+]]:vgpr(<2 x p1>), [[UV1:%[0-9]+]]:vgpr(<2 x p1>) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<4 x p1>)
   ; GREEDY-NEXT:   G_STORE [[UV]](<2 x p1>), [[DEF]](p1) :: (store (<2 x p1>) into `ptr addrspace(1) undef`, align 32, addrspace 1)
   ; GREEDY-NEXT:   [[C2:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 16
   ; GREEDY-NEXT:   [[PTR_ADD:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C2]](s64)
   ; GREEDY-NEXT:   G_STORE [[UV1]](<2 x p1>), [[PTR_ADD]](p1) :: (store (<2 x p1>) into `ptr addrspace(1) undef` + 16, basealign 32, addrspace 1)
   ; GREEDY-NEXT:   S_ENDPGM 0
   %val = call <4 x ptr addrspace(1)> @llvm.amdgcn.s.buffer.load.v4p1(<4 x i32> %rsrc, i32 %soffset, i32 0)
define amdgpu_ps void @s_buffer_load_v8p1_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[DEF:%[0-9]+]]:sgpr(p1) = G_IMPLICIT_DEF
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4, addrspace 7)
   ; CHECK-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x p1>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<2 x p1>), [[AMDGPU_BUFFER_LOAD1]](<2 x p1>), [[AMDGPU_BUFFER_LOAD2]](<2 x p1>), [[AMDGPU_BUFFER_LOAD3]](<2 x p1>)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:vgpr(<2 x p1>), [[UV1:%[0-9]+]]:vgpr(<2 x p1>), [[UV2:%[0-9]+]]:vgpr(<2 x p1>), [[UV3:%[0-9]+]]:vgpr(<2 x p1>) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<8 x p1>)
   ; CHECK-NEXT:   G_STORE [[UV]](<2 x p1>), [[DEF]](p1) :: (store (<2 x p1>) into `ptr addrspace(1) undef`, align 64, addrspace 1)
   ; CHECK-NEXT:   [[C2:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 16
   ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C2]](s64)
   ; CHECK-NEXT:   G_STORE [[UV1]](<2 x p1>), [[PTR_ADD]](p1) :: (store (<2 x p1>) into `ptr addrspace(1) undef` + 16, basealign 64, addrspace 1)
   ; CHECK-NEXT:   [[C3:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 32
   ; CHECK-NEXT:   [[PTR_ADD1:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C3]](s64)
define amdgpu_ps void @s_buffer_load_v8p1_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[DEF:%[0-9]+]]:sgpr(p1) = G_IMPLICIT_DEF
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4, addrspace 7)
   ; GREEDY-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x p1>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<2 x p1>), [[AMDGPU_BUFFER_LOAD1]](<2 x p1>), [[AMDGPU_BUFFER_LOAD2]](<2 x p1>), [[AMDGPU_BUFFER_LOAD3]](<2 x p1>)
   ; GREEDY-NEXT:   [[UV:%[0-9]+]]:vgpr(<2 x p1>), [[UV1:%[0-9]+]]:vgpr(<2 x p1>), [[UV2:%[0-9]+]]:vgpr(<2 x p1>), [[UV3:%[0-9]+]]:vgpr(<2 x p1>) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<8 x p1>)
   ; GREEDY-NEXT:   G_STORE [[UV]](<2 x p1>), [[DEF]](p1) :: (store (<2 x p1>) into `ptr addrspace(1) undef`, align 64, addrspace 1)
   ; GREEDY-NEXT:   [[C2:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 16
   ; GREEDY-NEXT:   [[PTR_ADD:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C2]](s64)
   ; GREEDY-NEXT:   G_STORE [[UV1]](<2 x p1>), [[PTR_ADD]](p1) :: (store (<2 x p1>) into `ptr addrspace(1) undef` + 16, basealign 64, addrspace 1)
   ; GREEDY-NEXT:   [[C3:%[0-9]+]]:sgpr(s64) = G_CONSTANT i64 32
   ; GREEDY-NEXT:   [[PTR_ADD1:%[0-9]+]]:sgpr(p1) = G_PTR_ADD [[DEF]], [[C3]](s64)
define amdgpu_ps float @s_buffer_load_f32_vgpr_offset_add_4092(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 4092
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[C]](s32)
   ; CHECK-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY4]], [[COPY5]]
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   [[C2:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4092, 0, 0 :: (dereferenceable invariant load (s32))
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4092, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; CHECK-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
   ; CHECK-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GREEDY-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4092
   ; GREEDY: bb.1 (%ir-block.0):
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 4092
   ; GREEDY-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[C]](s32)
   ; GREEDY-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY4]], [[COPY5]]
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; GREEDY-NEXT:   [[C2:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4092, 0, 0 :: (dereferenceable invariant load (s32))
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4092, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GREEDY-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
   ; GREEDY-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   %soffset = add i32 %soffset.base, 4092
   %val = call float @llvm.amdgcn.s.buffer.load.f32(<4 x i32> %rsrc, i32 %soffset, i32 0)
   ret float %val
 }
 
 define amdgpu_ps float @s_buffer_load_f32_vgpr_offset_add_4095(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; CHECK-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4095
   ; CHECK: bb.1 (%ir-block.0):
   ; CHECK-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 4095
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[C]](s32)
   ; CHECK-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY4]], [[COPY5]]
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   [[C2:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4095, 0, 0 :: (dereferenceable invariant load (s32))
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4095, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; CHECK-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
   ; CHECK-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GREEDY-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4095
   ; GREEDY: bb.1 (%ir-block.0):
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 4095
   ; GREEDY-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[C]](s32)
   ; GREEDY-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY4]], [[COPY5]]
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; GREEDY-NEXT:   [[C2:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4095, 0, 0 :: (dereferenceable invariant load (s32))
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4095, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GREEDY-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
   ; GREEDY-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   %soffset = add i32 %soffset.base, 4095
   %val = call float @llvm.amdgcn.s.buffer.load.f32(<4 x i32> %rsrc, i32 %soffset, i32 0)
   ret float %val
 }
 
 define amdgpu_ps float @s_buffer_load_f32_vgpr_offset_add_4096(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; CHECK-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4096
   ; CHECK: bb.1 (%ir-block.0):
   ; CHECK-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 4096
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[C]](s32)
   ; CHECK-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY4]], [[COPY5]]
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s32))
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; CHECK-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
   ; CHECK-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GREEDY-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4096
   ; GREEDY: bb.1 (%ir-block.0):
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 4096
   ; GREEDY-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[C]](s32)
   ; GREEDY-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY4]], [[COPY5]]
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s32))
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GREEDY-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
   ; GREEDY-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   %soffset = add i32 %soffset.base, 4096
   %val = call float @llvm.amdgcn.s.buffer.load.f32(<4 x i32> %rsrc, i32 %soffset, i32 0)
   ret float %val
 }
 
 ; Make sure the base offset is added to each split load.
 define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_add_4064(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; CHECK-LABEL: name: s_buffer_load_v8f32_vgpr_offset_add_4064
   ; CHECK: bb.1 (%ir-block.0):
   ; CHECK-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 4064
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[C]](s32)
   ; CHECK-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY4]], [[COPY5]]
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   [[C2:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4064, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4080, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4064, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4080, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; CHECK-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:vgpr(s32), [[UV1:%[0-9]+]]:vgpr(s32), [[UV2:%[0-9]+]]:vgpr(s32), [[UV3:%[0-9]+]]:vgpr(s32), [[UV4:%[0-9]+]]:vgpr(s32), [[UV5:%[0-9]+]]:vgpr(s32), [[UV6:%[0-9]+]]:vgpr(s32), [[UV7:%[0-9]+]]:vgpr(s32) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<8 x s32>)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
   ; CHECK-NEXT:   $vgpr4 = COPY [[UV4]](s32)
   ; CHECK-NEXT:   $vgpr5 = COPY [[UV5]](s32)
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_add_4064(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 4064
   ; GREEDY-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[C]](s32)
   ; GREEDY-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY4]], [[COPY5]]
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; GREEDY-NEXT:   [[C2:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4064, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4080, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4064, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4080, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GREEDY-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>)
   ; GREEDY-NEXT:   [[UV:%[0-9]+]]:vgpr(s32), [[UV1:%[0-9]+]]:vgpr(s32), [[UV2:%[0-9]+]]:vgpr(s32), [[UV3:%[0-9]+]]:vgpr(s32), [[UV4:%[0-9]+]]:vgpr(s32), [[UV5:%[0-9]+]]:vgpr(s32), [[UV6:%[0-9]+]]:vgpr(s32), [[UV7:%[0-9]+]]:vgpr(s32) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<8 x s32>)
   ; GREEDY-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; GREEDY-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; GREEDY-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; GREEDY-NEXT:   $vgpr3 = COPY [[UV3]](s32)
   ; GREEDY-NEXT:   $vgpr4 = COPY [[UV4]](s32)
   ; GREEDY-NEXT:   $vgpr5 = COPY [[UV5]](s32)
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_add_4068(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 4068
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[C]](s32)
   ; CHECK-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY4]], [[COPY5]]
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; CHECK-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:vgpr(s32), [[UV1:%[0-9]+]]:vgpr(s32), [[UV2:%[0-9]+]]:vgpr(s32), [[UV3:%[0-9]+]]:vgpr(s32), [[UV4:%[0-9]+]]:vgpr(s32), [[UV5:%[0-9]+]]:vgpr(s32), [[UV6:%[0-9]+]]:vgpr(s32), [[UV7:%[0-9]+]]:vgpr(s32) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<8 x s32>)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
   ; CHECK-NEXT:   $vgpr4 = COPY [[UV4]](s32)
   ; CHECK-NEXT:   $vgpr5 = COPY [[UV5]](s32)
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_add_4068(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 4068
   ; GREEDY-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[C]](s32)
   ; GREEDY-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY4]], [[COPY5]]
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GREEDY-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>)
   ; GREEDY-NEXT:   [[UV:%[0-9]+]]:vgpr(s32), [[UV1:%[0-9]+]]:vgpr(s32), [[UV2:%[0-9]+]]:vgpr(s32), [[UV3:%[0-9]+]]:vgpr(s32), [[UV4:%[0-9]+]]:vgpr(s32), [[UV5:%[0-9]+]]:vgpr(s32), [[UV6:%[0-9]+]]:vgpr(s32), [[UV7:%[0-9]+]]:vgpr(s32) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<8 x s32>)
   ; GREEDY-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; GREEDY-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; GREEDY-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; GREEDY-NEXT:   $vgpr3 = COPY [[UV3]](s32)
   ; GREEDY-NEXT:   $vgpr4 = COPY [[UV4]](s32)
   ; GREEDY-NEXT:   $vgpr5 = COPY [[UV5]](s32)
define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset_add_4032(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 4032
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[C]](s32)
   ; CHECK-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY4]], [[COPY5]]
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   [[C2:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4032, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4048, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4064, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4080, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4032, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4048, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4064, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4080, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4, addrspace 7)
   ; CHECK-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<16 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>), [[AMDGPU_BUFFER_LOAD2]](<4 x s32>), [[AMDGPU_BUFFER_LOAD3]](<4 x s32>)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:vgpr(s32), [[UV1:%[0-9]+]]:vgpr(s32), [[UV2:%[0-9]+]]:vgpr(s32), [[UV3:%[0-9]+]]:vgpr(s32), [[UV4:%[0-9]+]]:vgpr(s32), [[UV5:%[0-9]+]]:vgpr(s32), [[UV6:%[0-9]+]]:vgpr(s32), [[UV7:%[0-9]+]]:vgpr(s32), [[UV8:%[0-9]+]]:vgpr(s32), [[UV9:%[0-9]+]]:vgpr(s32), [[UV10:%[0-9]+]]:vgpr(s32), [[UV11:%[0-9]+]]:vgpr(s32), [[UV12:%[0-9]+]]:vgpr(s32), [[UV13:%[0-9]+]]:vgpr(s32), [[UV14:%[0-9]+]]:vgpr(s32), [[UV15:%[0-9]+]]:vgpr(s32) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<16 x s32>)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
   ; CHECK-NEXT:   $vgpr4 = COPY [[UV4]](s32)
   ; CHECK-NEXT:   $vgpr5 = COPY [[UV5]](s32)
define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset_add_4032(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 4032
   ; GREEDY-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[C]](s32)
   ; GREEDY-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY4]], [[COPY5]]
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; GREEDY-NEXT:   [[C2:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4032, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4048, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4064, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4080, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4032, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4048, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4064, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4080, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4, addrspace 7)
   ; GREEDY-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<16 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>), [[AMDGPU_BUFFER_LOAD2]](<4 x s32>), [[AMDGPU_BUFFER_LOAD3]](<4 x s32>)
   ; GREEDY-NEXT:   [[UV:%[0-9]+]]:vgpr(s32), [[UV1:%[0-9]+]]:vgpr(s32), [[UV2:%[0-9]+]]:vgpr(s32), [[UV3:%[0-9]+]]:vgpr(s32), [[UV4:%[0-9]+]]:vgpr(s32), [[UV5:%[0-9]+]]:vgpr(s32), [[UV6:%[0-9]+]]:vgpr(s32), [[UV7:%[0-9]+]]:vgpr(s32), [[UV8:%[0-9]+]]:vgpr(s32), [[UV9:%[0-9]+]]:vgpr(s32), [[UV10:%[0-9]+]]:vgpr(s32), [[UV11:%[0-9]+]]:vgpr(s32), [[UV12:%[0-9]+]]:vgpr(s32), [[UV13:%[0-9]+]]:vgpr(s32), [[UV14:%[0-9]+]]:vgpr(s32), [[UV15:%[0-9]+]]:vgpr(s32) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<16 x s32>)
   ; GREEDY-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; GREEDY-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; GREEDY-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; GREEDY-NEXT:   $vgpr3 = COPY [[UV3]](s32)
   ; GREEDY-NEXT:   $vgpr4 = COPY [[UV4]](s32)
   ; GREEDY-NEXT:   $vgpr5 = COPY [[UV5]](s32)
define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset_add_4036(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 4036
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[C]](s32)
   ; CHECK-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY4]], [[COPY5]]
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4, addrspace 7)
   ; CHECK-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<16 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>), [[AMDGPU_BUFFER_LOAD2]](<4 x s32>), [[AMDGPU_BUFFER_LOAD3]](<4 x s32>)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:vgpr(s32), [[UV1:%[0-9]+]]:vgpr(s32), [[UV2:%[0-9]+]]:vgpr(s32), [[UV3:%[0-9]+]]:vgpr(s32), [[UV4:%[0-9]+]]:vgpr(s32), [[UV5:%[0-9]+]]:vgpr(s32), [[UV6:%[0-9]+]]:vgpr(s32), [[UV7:%[0-9]+]]:vgpr(s32), [[UV8:%[0-9]+]]:vgpr(s32), [[UV9:%[0-9]+]]:vgpr(s32), [[UV10:%[0-9]+]]:vgpr(s32), [[UV11:%[0-9]+]]:vgpr(s32), [[UV12:%[0-9]+]]:vgpr(s32), [[UV13:%[0-9]+]]:vgpr(s32), [[UV14:%[0-9]+]]:vgpr(s32), [[UV15:%[0-9]+]]:vgpr(s32) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<16 x s32>)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
   ; CHECK-NEXT:   $vgpr4 = COPY [[UV4]](s32)
   ; CHECK-NEXT:   $vgpr5 = COPY [[UV5]](s32)
define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset_add_4036(<4 x i32> inreg %rsrc, i32 %soffset.base) {
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 4036
   ; GREEDY-NEXT:   [[COPY5:%[0-9]+]]:vgpr(s32) = COPY [[C]](s32)
   ; GREEDY-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY4]], [[COPY5]]
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 16, align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 48, align 4, addrspace 7)
   ; GREEDY-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<16 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>), [[AMDGPU_BUFFER_LOAD2]](<4 x s32>), [[AMDGPU_BUFFER_LOAD3]](<4 x s32>)
   ; GREEDY-NEXT:   [[UV:%[0-9]+]]:vgpr(s32), [[UV1:%[0-9]+]]:vgpr(s32), [[UV2:%[0-9]+]]:vgpr(s32), [[UV3:%[0-9]+]]:vgpr(s32), [[UV4:%[0-9]+]]:vgpr(s32), [[UV5:%[0-9]+]]:vgpr(s32), [[UV6:%[0-9]+]]:vgpr(s32), [[UV7:%[0-9]+]]:vgpr(s32), [[UV8:%[0-9]+]]:vgpr(s32), [[UV9:%[0-9]+]]:vgpr(s32), [[UV10:%[0-9]+]]:vgpr(s32), [[UV11:%[0-9]+]]:vgpr(s32), [[UV12:%[0-9]+]]:vgpr(s32), [[UV13:%[0-9]+]]:vgpr(s32), [[UV14:%[0-9]+]]:vgpr(s32), [[UV15:%[0-9]+]]:vgpr(s32) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<16 x s32>)
   ; GREEDY-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; GREEDY-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; GREEDY-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; GREEDY-NEXT:   $vgpr3 = COPY [[UV3]](s32)
   ; GREEDY-NEXT:   $vgpr4 = COPY [[UV4]](s32)
   ; GREEDY-NEXT:   $vgpr5 = COPY [[UV5]](s32)
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc(<4 x i32> %rsrc, i32 inreg %soffset) {
   ; CHECK-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; CHECK-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; CHECK-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; CHECK-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.3:
   ; CHECK-NEXT:   successors: %bb.4, %bb.2
   ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY5]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s32))
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY5]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; CHECK-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; CHECK-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.4:
   ; CHECK-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.5:
   ; CHECK-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc(<4 x i32> %rsrc, i32 inreg %soffset) {
   ; GREEDY-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; GREEDY-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; GREEDY-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; GREEDY-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.3:
   ; GREEDY-NEXT:   successors: %bb.4, %bb.2
   ; GREEDY-NEXT: {{  $}}
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY5]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s32))
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY5]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GREEDY-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GREEDY-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.4:
   ; GREEDY-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.5:
   ; GREEDY-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_soffset_add_4092(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; CHECK-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; CHECK-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; CHECK-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; CHECK-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.3:
   ; CHECK-NEXT:   successors: %bb.4, %bb.2
   ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[C1]], [[COPY4]], 4092, 0, 0 :: (dereferenceable invariant load (s32))
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[C1]], [[COPY4]], 4092, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; CHECK-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; CHECK-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.4:
   ; CHECK-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.5:
   ; CHECK-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_soffset_add_4092(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; GREEDY-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; GREEDY-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; GREEDY-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; GREEDY-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.3:
   ; GREEDY-NEXT:   successors: %bb.4, %bb.2
   ; GREEDY-NEXT: {{  $}}
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[C1]], [[COPY4]], 4092, 0, 0 :: (dereferenceable invariant load (s32))
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[C1]], [[COPY4]], 4092, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GREEDY-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GREEDY-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.4:
   ; GREEDY-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.5:
   ; GREEDY-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_soffset_add_4096(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; CHECK-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; CHECK-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; CHECK-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; CHECK-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.3:
   ; CHECK-NEXT:   successors: %bb.4, %bb.2
   ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY5]], [[C1]], 0, 0, 0 :: (dereferenceable invariant load (s32))
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY5]], [[C1]], 0, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; CHECK-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; CHECK-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.4:
   ; CHECK-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.5:
   ; CHECK-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_soffset_add_4096(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; GREEDY-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; GREEDY-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; GREEDY-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; GREEDY-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.3:
   ; GREEDY-NEXT:   successors: %bb.4, %bb.2
   ; GREEDY-NEXT: {{  $}}
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY5]], [[C1]], 0, 0, 0 :: (dereferenceable invariant load (s32))
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY5]], [[C1]], 0, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GREEDY-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GREEDY-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.4:
   ; GREEDY-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.5:
   ; GREEDY-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_offset_4095(<4 x i32> %rsrc) {
   ; CHECK-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; CHECK-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; CHECK-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; CHECK-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.3:
   ; CHECK-NEXT:   successors: %bb.4, %bb.2
   ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C3]](s32), [[C1]], [[C2]], 4095, 0, 0 :: (dereferenceable invariant load (s32) from unknown-address + 4095, align 1)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C3]](s32), [[C1]], [[C2]], 4095, 0, 0 :: (dereferenceable invariant load (s32) from unknown-address + 4095, align 1, addrspace 7)
   ; CHECK-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; CHECK-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.4:
   ; CHECK-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.5:
   ; CHECK-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_offset_4095(<4 x i32> %rsrc) {
   ; GREEDY-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; GREEDY-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; GREEDY-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; GREEDY-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.3:
   ; GREEDY-NEXT:   successors: %bb.4, %bb.2
   ; GREEDY-NEXT: {{  $}}
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C3]](s32), [[C1]], [[C2]], 4095, 0, 0 :: (dereferenceable invariant load (s32) from unknown-address + 4095, align 1)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C3]](s32), [[C1]], [[C2]], 4095, 0, 0 :: (dereferenceable invariant load (s32) from unknown-address + 4095, align 1, addrspace 7)
   ; GREEDY-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GREEDY-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.4:
   ; GREEDY-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.5:
   ; GREEDY-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_offset_4096(<4 x i32> %rsrc) {
   ; CHECK-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; CHECK-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; CHECK-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; CHECK-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.3:
   ; CHECK-NEXT:   successors: %bb.4, %bb.2
   ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 0, 0, 0 :: (dereferenceable invariant load (s32))
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 0, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; CHECK-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; CHECK-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.4:
   ; CHECK-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.5:
   ; CHECK-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_offset_4096(<4 x i32> %rsrc) {
   ; GREEDY-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; GREEDY-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; GREEDY-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; GREEDY-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.3:
   ; GREEDY-NEXT:   successors: %bb.4, %bb.2
   ; GREEDY-NEXT: {{  $}}
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 0, 0, 0 :: (dereferenceable invariant load (s32))
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 0, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GREEDY-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GREEDY-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.4:
   ; GREEDY-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.5:
   ; GREEDY-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4064(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; CHECK-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; CHECK-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; CHECK-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; CHECK-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.3:
   ; CHECK-NEXT:   successors: %bb.4, %bb.2
   ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[C1]], [[COPY4]], 4064, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[C1]], [[COPY4]], 4080, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[C1]], [[COPY4]], 4064, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[C1]], [[COPY4]], 4080, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; CHECK-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; CHECK-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.4:
   ; CHECK-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.5:
   ; CHECK-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>)
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4064(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; GREEDY-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; GREEDY-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; GREEDY-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; GREEDY-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.3:
   ; GREEDY-NEXT:   successors: %bb.4, %bb.2
   ; GREEDY-NEXT: {{  $}}
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[C1]], [[COPY4]], 4064, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[C1]], [[COPY4]], 4080, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[C1]], [[COPY4]], 4064, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[C1]], [[COPY4]], 4080, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GREEDY-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GREEDY-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.4:
   ; GREEDY-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.5:
   ; GREEDY-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>)
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4068(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; CHECK-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; CHECK-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; CHECK-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; CHECK-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.3:
   ; CHECK-NEXT:   successors: %bb.4, %bb.2
   ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY5]], [[C1]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY5]], [[C1]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY5]], [[C1]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY5]], [[C1]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; CHECK-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; CHECK-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.4:
   ; CHECK-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.5:
   ; CHECK-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>)
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4068(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; GREEDY-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; GREEDY-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; GREEDY-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; GREEDY-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.3:
   ; GREEDY-NEXT:   successors: %bb.4, %bb.2
   ; GREEDY-NEXT: {{  $}}
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY5]], [[C1]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY5]], [[C1]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY5]], [[C1]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY5]], [[C1]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GREEDY-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GREEDY-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.4:
   ; GREEDY-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.5:
   ; GREEDY-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>)
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4096(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; CHECK-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; CHECK-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; CHECK-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; CHECK-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.3:
   ; CHECK-NEXT:   successors: %bb.4, %bb.2
   ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY5]], [[C1]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY5]], [[C1]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY5]], [[C1]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY5]], [[C1]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; CHECK-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; CHECK-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.4:
   ; CHECK-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.5:
   ; CHECK-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>)
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4096(<4 x i32> %rsrc, i32 inreg %soffset.base) {
   ; GREEDY-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; GREEDY-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; GREEDY-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; GREEDY-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.3:
   ; GREEDY-NEXT:   successors: %bb.4, %bb.2
   ; GREEDY-NEXT: {{  $}}
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY5]], [[C1]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY5]], [[C1]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY5]], [[C1]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C2]](s32), [[COPY5]], [[C1]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GREEDY-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GREEDY-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.4:
   ; GREEDY-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.5:
   ; GREEDY-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>)
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_5000(<4 x i32> %rsrc, i32 %offset.base) {
   ; CHECK-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; CHECK-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; CHECK-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; CHECK-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.3:
   ; CHECK-NEXT:   successors: %bb.4, %bb.2
   ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; CHECK-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; CHECK-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.4:
   ; CHECK-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.5:
   ; CHECK-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>)
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_5000(<4 x i32> %rsrc, i32 %offset.base) {
   ; GREEDY-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; GREEDY-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; GREEDY-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; GREEDY-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.3:
   ; GREEDY-NEXT:   successors: %bb.4, %bb.2
   ; GREEDY-NEXT: {{  $}}
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GREEDY-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GREEDY-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.4:
   ; GREEDY-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.5:
   ; GREEDY-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>)
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4076(<4 x i32> %rsrc, i32 %offset.base) {
   ; CHECK-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; CHECK-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; CHECK-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; CHECK-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.3:
   ; CHECK-NEXT:   successors: %bb.4, %bb.2
   ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; CHECK-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; CHECK-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.4:
   ; CHECK-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.5:
   ; CHECK-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>)
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4076(<4 x i32> %rsrc, i32 %offset.base) {
   ; GREEDY-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; GREEDY-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; GREEDY-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; GREEDY-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.3:
   ; GREEDY-NEXT:   successors: %bb.4, %bb.2
   ; GREEDY-NEXT: {{  $}}
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GREEDY-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GREEDY-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.4:
   ; GREEDY-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.5:
   ; GREEDY-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>)
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4080(<4 x i32> %rsrc, i32 %offset.base) {
   ; CHECK-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; CHECK-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; CHECK-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; CHECK-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.3:
   ; CHECK-NEXT:   successors: %bb.4, %bb.2
   ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; CHECK-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; CHECK-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.4:
   ; CHECK-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.5:
   ; CHECK-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>)
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4080(<4 x i32> %rsrc, i32 %offset.base) {
   ; GREEDY-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; GREEDY-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; GREEDY-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; GREEDY-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.3:
   ; GREEDY-NEXT:   successors: %bb.4, %bb.2
   ; GREEDY-NEXT: {{  $}}
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load (s128), align 4, addrspace 7)
   ; GREEDY-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GREEDY-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.4:
   ; GREEDY-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.5:
   ; GREEDY-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>)
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_offset_4064(<4 x i32> %rsrc, i32 %offset.base) {
   ; CHECK-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; CHECK-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; CHECK-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; CHECK-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.3:
   ; CHECK-NEXT:   successors: %bb.4, %bb.2
   ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C3]](s32), [[C1]], [[C2]], 4064, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4)
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C3]](s32), [[C1]], [[C2]], 4080, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C3]](s32), [[C1]], [[C2]], 4064, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4, addrspace 7)
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C3]](s32), [[C1]], [[C2]], 4080, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4, addrspace 7)
   ; CHECK-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; CHECK-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.4:
   ; CHECK-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.5:
   ; CHECK-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>)
define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_offset_4064(<4 x i32> %rsrc, i32 %offset.base) {
   ; GREEDY-NEXT:   [[ICMP1:%[0-9]+]]:vcc(s1) = G_ICMP intpred(eq), [[UV7]](s64), [[UV5]]
   ; GREEDY-NEXT:   [[AND:%[0-9]+]]:vcc(s1) = G_AND [[ICMP]], [[ICMP1]]
   ; GREEDY-NEXT:   [[INT:%[0-9]+]]:sreg_64_xexec(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.ballot), [[AND]](s1)
   ; GREEDY-NEXT:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[INT]](s64), implicit-def $exec, implicit-def $scc, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.3:
   ; GREEDY-NEXT:   successors: %bb.4, %bb.2
   ; GREEDY-NEXT: {{  $}}
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C3]](s32), [[C1]], [[C2]], 4064, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4)
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C3]](s32), [[C1]], [[C2]], 4080, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C3]](s32), [[C1]], [[C2]], 4064, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4, addrspace 7)
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C3]](s32), [[C1]], [[C2]], 4080, 0, 0 :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4, addrspace 7)
   ; GREEDY-NEXT:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
   ; GREEDY-NEXT:   SI_WATERFALL_LOOP %bb.2, implicit $exec
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.4:
   ; GREEDY-NEXT:   $exec = S_MOV_B64_term [[S_MOV_B64_]]
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT: bb.5:
   ; GREEDY-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>)
define amdgpu_ps float @s_buffer_load_f32_offset_add_vgpr_sgpr(<4 x i32> inreg %rsrc, i32 %offset.v, i32 inreg %offset.s) {
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:sgpr(s32) = COPY $sgpr6
   ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:vgpr(s32) = COPY [[COPY5]](s32)
   ; CHECK-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY4]], [[COPY6]]
   ; CHECK-NEXT:   [[C:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C]](s32), [[COPY4]], [[COPY5]], 0, 0, 0 :: (dereferenceable invariant load (s32))
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C]](s32), [[COPY4]], [[COPY5]], 0, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; CHECK-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
   ; CHECK-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GREEDY-LABEL: name: s_buffer_load_f32_offset_add_vgpr_sgpr
   ; GREEDY: bb.1 (%ir-block.0):
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[COPY5:%[0-9]+]]:sgpr(s32) = COPY $sgpr6
   ; GREEDY-NEXT:   [[COPY6:%[0-9]+]]:vgpr(s32) = COPY [[COPY5]](s32)
   ; GREEDY-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY4]], [[COPY6]]
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C]](s32), [[COPY4]], [[COPY5]], 0, 0, 0 :: (dereferenceable invariant load (s32))
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C]](s32), [[COPY4]], [[COPY5]], 0, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GREEDY-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
   ; GREEDY-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   %offset = add i32 %offset.v, %offset.s
   %val = call float @llvm.amdgcn.s.buffer.load.f32(<4 x i32> %rsrc, i32 %offset, i32 0)
   ret float %val
 }
 
 define amdgpu_ps float @s_buffer_load_f32_offset_add_sgpr_vgpr(<4 x i32> inreg %rsrc, i32 %offset.v, i32 inreg %offset.s) {
   ; CHECK-LABEL: name: s_buffer_load_f32_offset_add_sgpr_vgpr
   ; CHECK: bb.1 (%ir-block.0):
   ; CHECK-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:sgpr(s32) = COPY $sgpr6
   ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:vgpr(s32) = COPY [[COPY5]](s32)
   ; CHECK-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY6]], [[COPY4]]
   ; CHECK-NEXT:   [[C:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C]](s32), [[COPY4]], [[COPY5]], 0, 0, 0 :: (dereferenceable invariant load (s32))
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C]](s32), [[COPY4]], [[COPY5]], 0, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; CHECK-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
   ; CHECK-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GREEDY-LABEL: name: s_buffer_load_f32_offset_add_sgpr_vgpr
   ; GREEDY: bb.1 (%ir-block.0):
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[COPY5:%[0-9]+]]:sgpr(s32) = COPY $sgpr6
   ; GREEDY-NEXT:   [[COPY6:%[0-9]+]]:vgpr(s32) = COPY [[COPY5]](s32)
   ; GREEDY-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY6]], [[COPY4]]
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C]](s32), [[COPY4]], [[COPY5]], 0, 0, 0 :: (dereferenceable invariant load (s32))
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C]](s32), [[COPY4]], [[COPY5]], 0, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GREEDY-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
   ; GREEDY-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   %offset = add i32 %offset.s, %offset.v
   %val = call float @llvm.amdgcn.s.buffer.load.f32(<4 x i32> %rsrc, i32 %offset, i32 0)
   ret float %val
 }
 
 define amdgpu_ps float @s_buffer_load_f32_offset_add_vgpr_sgpr_imm(<4 x i32> inreg %rsrc, i32 %offset.v, i32 inreg %offset.s) {
define amdgpu_ps float @s_buffer_load_f32_offset_add_vgpr_sgpr_imm(<4 x i32> inreg %rsrc, i32 %offset.v, i32 inreg %offset.s) {
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:sgpr(s32) = COPY $sgpr6
   ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:vgpr(s32) = COPY [[COPY5]](s32)
   ; CHECK-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY4]], [[COPY6]]
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 1024
   ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:vgpr(s32) = COPY [[C]](s32)
   ; CHECK-NEXT:   [[ADD1:%[0-9]+]]:vgpr(s32) = G_ADD [[ADD]], [[COPY7]]
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   [[C2:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[ADD]], [[C1]], 1024, 0, 0 :: (dereferenceable invariant load (s32))
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[ADD]], [[C1]], 1024, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; CHECK-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
   ; CHECK-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GREEDY-LABEL: name: s_buffer_load_f32_offset_add_vgpr_sgpr_imm
   ; GREEDY: bb.1 (%ir-block.0):
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[COPY5:%[0-9]+]]:sgpr(s32) = COPY $sgpr6
   ; GREEDY-NEXT:   [[COPY6:%[0-9]+]]:vgpr(s32) = COPY [[COPY5]](s32)
   ; GREEDY-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY4]], [[COPY6]]
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 1024
   ; GREEDY-NEXT:   [[COPY7:%[0-9]+]]:vgpr(s32) = COPY [[C]](s32)
   ; GREEDY-NEXT:   [[ADD1:%[0-9]+]]:vgpr(s32) = G_ADD [[ADD]], [[COPY7]]
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; GREEDY-NEXT:   [[C2:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[ADD]], [[C1]], 1024, 0, 0 :: (dereferenceable invariant load (s32))
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[ADD]], [[C1]], 1024, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GREEDY-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
   ; GREEDY-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   %offset.base = add i32 %offset.v, %offset.s
   %offset = add i32 %offset.base, 1024
   %val = call float @llvm.amdgcn.s.buffer.load.f32(<4 x i32> %rsrc, i32 %offset, i32 0)
   ret float %val
 }
 
define amdgpu_ps float @s_buffer_load_f32_offset_add_sgpr_vgpr_imm(<4 x i32> inreg %rsrc, i32 %offset.v, i32 inreg %offset.s) {
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:sgpr(s32) = COPY $sgpr6
   ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:vgpr(s32) = COPY [[COPY5]](s32)
   ; CHECK-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY6]], [[COPY4]]
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 1024
   ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:vgpr(s32) = COPY [[C]](s32)
   ; CHECK-NEXT:   [[ADD1:%[0-9]+]]:vgpr(s32) = G_ADD [[ADD]], [[COPY7]]
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   [[C2:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[ADD]], [[C1]], 1024, 0, 0 :: (dereferenceable invariant load (s32))
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[ADD]], [[C1]], 1024, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; CHECK-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
   ; CHECK-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GREEDY-LABEL: name: s_buffer_load_f32_offset_add_sgpr_vgpr_imm
   ; GREEDY: bb.1 (%ir-block.0):
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[COPY5:%[0-9]+]]:sgpr(s32) = COPY $sgpr6
   ; GREEDY-NEXT:   [[COPY6:%[0-9]+]]:vgpr(s32) = COPY [[COPY5]](s32)
   ; GREEDY-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY6]], [[COPY4]]
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 1024
   ; GREEDY-NEXT:   [[COPY7:%[0-9]+]]:vgpr(s32) = COPY [[C]](s32)
   ; GREEDY-NEXT:   [[ADD1:%[0-9]+]]:vgpr(s32) = G_ADD [[ADD]], [[COPY7]]
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
   ; GREEDY-NEXT:   [[C2:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[ADD]], [[C1]], 1024, 0, 0 :: (dereferenceable invariant load (s32))
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[ADD]], [[C1]], 1024, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GREEDY-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
   ; GREEDY-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   %offset.base = add i32 %offset.s, %offset.v
   %offset = add i32 %offset.base, 1024
   %val = call float @llvm.amdgcn.s.buffer.load.f32(<4 x i32> %rsrc, i32 %offset, i32 0)
   ret float %val
 }
 
define amdgpu_ps float @s_buffer_load_f32_offset_add_imm_sgpr_vgpr(<4 x i32> inreg %rsrc, i32 %offset.v, i32 inreg %offset.s) {
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:sgpr(s32) = COPY $sgpr6
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 1024
   ; CHECK-NEXT:   [[ADD:%[0-9]+]]:sgpr(s32) = G_ADD [[COPY5]], [[C]]
   ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:vgpr(s32) = COPY [[ADD]](s32)
   ; CHECK-NEXT:   [[ADD1:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY6]], [[COPY4]]
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[ADD]], 0, 0, 0 :: (dereferenceable invariant load (s32))
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[ADD]], 0, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; CHECK-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
   ; CHECK-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GREEDY-LABEL: name: s_buffer_load_f32_offset_add_imm_sgpr_vgpr
   ; GREEDY: bb.1 (%ir-block.0):
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[COPY5:%[0-9]+]]:sgpr(s32) = COPY $sgpr6
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 1024
   ; GREEDY-NEXT:   [[ADD:%[0-9]+]]:sgpr(s32) = G_ADD [[COPY5]], [[C]]
   ; GREEDY-NEXT:   [[COPY6:%[0-9]+]]:vgpr(s32) = COPY [[ADD]](s32)
   ; GREEDY-NEXT:   [[ADD1:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY6]], [[COPY4]]
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[ADD]], 0, 0, 0 :: (dereferenceable invariant load (s32))
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[ADD]], 0, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GREEDY-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
   ; GREEDY-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   %offset.base = add i32 %offset.s, 1024
   %offset = add i32 %offset.base, %offset.v
   %val = call float @llvm.amdgcn.s.buffer.load.f32(<4 x i32> %rsrc, i32 %offset, i32 0)
   ret float %val
 }
 
define amdgpu_ps float @s_buffer_load_f32_offset_add_imm_vgpr_sgpr(<4 x i32> inreg %rsrc, i32 %offset.v, i32 inreg %offset.s) {
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:sgpr(s32) = COPY $sgpr6
   ; CHECK-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 1024
   ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:vgpr(s32) = COPY [[C]](s32)
   ; CHECK-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY4]], [[COPY6]]
   ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:vgpr(s32) = COPY [[COPY5]](s32)
   ; CHECK-NEXT:   [[ADD1:%[0-9]+]]:vgpr(s32) = G_ADD [[ADD]], [[COPY7]]
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[ADD]], [[COPY5]], 0, 0, 0 :: (dereferenceable invariant load (s32))
+  ; CHECK-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[ADD]], [[COPY5]], 0, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; CHECK-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
   ; CHECK-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ; GREEDY-LABEL: name: s_buffer_load_f32_offset_add_imm_vgpr_sgpr
   ; GREEDY: bb.1 (%ir-block.0):
   ; GREEDY-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
   ; GREEDY-NEXT: {{  $}}
   ; GREEDY-NEXT:   [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr2
   ; GREEDY-NEXT:   [[COPY1:%[0-9]+]]:sgpr(s32) = COPY $sgpr3
   ; GREEDY-NEXT:   [[COPY2:%[0-9]+]]:sgpr(s32) = COPY $sgpr4
   ; GREEDY-NEXT:   [[COPY3:%[0-9]+]]:sgpr(s32) = COPY $sgpr5
   ; GREEDY-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
   ; GREEDY-NEXT:   [[COPY4:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
   ; GREEDY-NEXT:   [[COPY5:%[0-9]+]]:sgpr(s32) = COPY $sgpr6
   ; GREEDY-NEXT:   [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 1024
   ; GREEDY-NEXT:   [[COPY6:%[0-9]+]]:vgpr(s32) = COPY [[C]](s32)
   ; GREEDY-NEXT:   [[ADD:%[0-9]+]]:vgpr(s32) = G_ADD [[COPY4]], [[COPY6]]
   ; GREEDY-NEXT:   [[COPY7:%[0-9]+]]:vgpr(s32) = COPY [[COPY5]](s32)
   ; GREEDY-NEXT:   [[ADD1:%[0-9]+]]:vgpr(s32) = G_ADD [[ADD]], [[COPY7]]
   ; GREEDY-NEXT:   [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
-  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[ADD]], [[COPY5]], 0, 0, 0 :: (dereferenceable invariant load (s32))
+  ; GREEDY-NEXT:   [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[ADD]], [[COPY5]], 0, 0, 0 :: (dereferenceable invariant load (s32), addrspace 7)
   ; GREEDY-NEXT:   $vgpr0 = COPY [[AMDGPU_BUFFER_LOAD]](s32)
   ; GREEDY-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   %offset.base = add i32 %offset.v, 1024
   %offset = add i32 %offset.base, %offset.s
   %val = call float @llvm.amdgcn.s.buffer.load.f32(<4 x i32> %rsrc, i32 %offset, i32 0)
   ret float %val
 }
 

llvm/test/CodeGen/AMDGPU/GlobalISel/regbankselect-amdgcn.s.buffer.load.mir

bb.0:
     ; GREEDY-NEXT: [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
     ; GREEDY-NEXT: [[C2:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
     ; GREEDY-NEXT: [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[COPY]](<4 x s32>), [[C2]](s32), [[COPY2]], [[C1]], 256, 0, 0 :: (dereferenceable invariant load (s32))
     ; GREEDY-NEXT: S_ENDPGM 0, implicit [[AMDGPU_BUFFER_LOAD]](s32)
     %0:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
     %1:_(s32) = COPY $sgpr0
     %2:vgpr(s32) = G_CONSTANT i32 256
     %3:_(s32) = G_ADD %1, %2
-    %4:_(s32) = G_AMDGPU_S_BUFFER_LOAD %0, %3, 0
+    %4:_(s32) = G_AMDGPU_S_BUFFER_LOAD %0, %3, 0 :: (dereferenceable invariant load (s32))
     S_ENDPGM 0, implicit %4
 
 ...
 
 ---
 name:            s_buffer_load_negative_offset
 legalized:       true
 tracksRegLiveness: true
bb.0:
     ; GREEDY-NEXT: [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 0
     ; GREEDY-NEXT: [[C2:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
     ; GREEDY-NEXT: [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[COPY]](<4 x s32>), [[C2]](s32), [[ADD]], [[C1]], 0, 0, 0 :: (dereferenceable invariant load (s32))
     ; GREEDY-NEXT: S_ENDPGM 0, implicit [[AMDGPU_BUFFER_LOAD]](s32)
     %0:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
     %1:_(s32) = COPY $vgpr0
     %2:_(s32) = G_CONSTANT i32 -60
     %3:_(s32) = G_ADD %1, %2
-    %4:_(s32) = G_AMDGPU_S_BUFFER_LOAD %0, %3, 0
+    %4:_(s32) = G_AMDGPU_S_BUFFER_LOAD %0, %3, 0 :: (dereferenceable invariant load (s32))
     S_ENDPGM 0, implicit %4
 
 ...

llvm/test/CodeGen/AMDGPU/splitkit-getsubrangeformask.ll

define amdgpu_gs void @_amdgpu_gs_main(i32 inreg %primShaderTableAddrLow, <31 x i32> inreg %userData) {
   ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %130, 16, 0 :: (invariant load (s128) from %ir.81, addrspace 4)
   ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM1:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM undef %74:sreg_64, 0, 0 :: (invariant load (s128) from `ptr addrspace(4) undef`, addrspace 4)
   ; CHECK-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM undef %132:sgpr_128, 0, 0 :: (dereferenceable invariant load (s32))
   ; CHECK-NEXT:   KILL undef %74:sreg_64
   ; CHECK-NEXT:   KILL undef %132:sgpr_128
   ; CHECK-NEXT:   KILL %130.sub0, %130.sub1
   ; CHECK-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM1:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[S_LOAD_DWORDX4_IMM]], 0, 0 :: (dereferenceable invariant load (s32))
   ; CHECK-NEXT:   [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
-  ; CHECK-NEXT:   undef %302.sub1:sgpr_128 = S_MOV_B32 0
+  ; CHECK-NEXT:   undef %304.sub1:sgpr_128 = S_MOV_B32 0
   ; CHECK-NEXT:   [[BUFFER_LOAD_FORMAT_X_IDXEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_IDXEN [[V_MOV_B32_e32_]], undef %89:sgpr_128, 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
   ; CHECK-NEXT:   [[BUFFER_LOAD_FORMAT_X_IDXEN1:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_IDXEN [[V_MOV_B32_e32_]], [[S_LOAD_DWORDX4_IMM1]], 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
   ; CHECK-NEXT:   KILL undef %89:sgpr_128
   ; CHECK-NEXT:   [[S_SUB_I32_:%[0-9]+]]:sreg_32 = S_SUB_I32 [[S_BUFFER_LOAD_DWORD_IMM]], 29, implicit-def dead $scc
   ; CHECK-NEXT:   [[S_SUB_I32_1:%[0-9]+]]:sreg_32 = S_SUB_I32 [[S_BUFFER_LOAD_DWORD_IMM]], 30, implicit-def dead $scc
   ; CHECK-NEXT:   [[S_SUB_I32_2:%[0-9]+]]:sreg_32 = S_SUB_I32 [[S_BUFFER_LOAD_DWORD_IMM1]], 31, implicit-def dead $scc
   ; CHECK-NEXT:   [[S_ADD_U32_:%[0-9]+]]:sreg_32 = S_ADD_U32 [[COPY5]], 64, implicit-def $scc
   ; CHECK-NEXT:   [[S_ADDC_U32_:%[0-9]+]]:sreg_32 = S_ADDC_U32 undef %54:sreg_32, 0, implicit-def dead $scc, implicit $scc
-  ; CHECK-NEXT:   undef %149.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_]], [[S_LSHL_B32_]], implicit-def $scc
-  ; CHECK-NEXT:   %149.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_]], [[S_ASHR_I32_]], implicit-def dead $scc, implicit $scc
-  ; CHECK-NEXT:   undef %156.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_]], [[S_LSHL_B32_1]], implicit-def $scc
-  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM2:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %149, 0, 0 :: (invariant load (s128) from %ir.87, addrspace 4)
-  ; CHECK-NEXT:   %156.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_]], [[S_ASHR_I32_1]], implicit-def dead $scc, implicit $scc
-  ; CHECK-NEXT:   undef %163.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_]], [[S_LSHL_B32_2]], implicit-def $scc
-  ; CHECK-NEXT:   %163.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_]], [[S_ASHR_I32_2]], implicit-def dead $scc, implicit $scc
-  ; CHECK-NEXT:   [[S_ASHR_I32_3:%[0-9]+]]:sreg_32_xm0 = S_ASHR_I32 undef %171:sreg_32, 31, implicit-def dead $scc
-  ; CHECK-NEXT:   undef %176.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_]], undef %171:sreg_32, implicit-def $scc
-  ; CHECK-NEXT:   %176.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_]], [[S_ASHR_I32_3]], implicit-def dead $scc, implicit $scc
-  ; CHECK-NEXT:   undef %183.sub0:sreg_64 = S_ADD_U32 %50.sub0, [[S_LSHL_B32_]], implicit-def $scc
-  ; CHECK-NEXT:   %183.sub1:sreg_64 = S_ADDC_U32 undef %51:sreg_32, [[S_ASHR_I32_]], implicit-def dead $scc, implicit $scc
-  ; CHECK-NEXT:   undef %190.sub0:sreg_64 = S_ADD_U32 %50.sub0, [[S_LSHL_B32_1]], implicit-def $scc
-  ; CHECK-NEXT:   %190.sub1:sreg_64 = S_ADDC_U32 undef %51:sreg_32, [[S_ASHR_I32_1]], implicit-def dead $scc, implicit $scc
-  ; CHECK-NEXT:   undef %200.sub0:sreg_64 = S_ADD_U32 %50.sub0, undef %171:sreg_32, implicit-def $scc
-  ; CHECK-NEXT:   %200.sub1:sreg_64 = S_ADDC_U32 undef %51:sreg_32, [[S_ASHR_I32_3]], implicit-def dead $scc, implicit $scc
+  ; CHECK-NEXT:   undef %151.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_]], [[S_LSHL_B32_]], implicit-def $scc
+  ; CHECK-NEXT:   %151.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_]], [[S_ASHR_I32_]], implicit-def dead $scc, implicit $scc
+  ; CHECK-NEXT:   undef %158.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_]], [[S_LSHL_B32_1]], implicit-def $scc
+  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM2:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %151, 0, 0 :: (invariant load (s128) from %ir.87, addrspace 4)
+  ; CHECK-NEXT:   %158.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_]], [[S_ASHR_I32_1]], implicit-def dead $scc, implicit $scc
+  ; CHECK-NEXT:   undef %165.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_]], [[S_LSHL_B32_2]], implicit-def $scc
+  ; CHECK-NEXT:   %165.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_]], [[S_ASHR_I32_2]], implicit-def dead $scc, implicit $scc
+  ; CHECK-NEXT:   [[S_ASHR_I32_3:%[0-9]+]]:sreg_32_xm0 = S_ASHR_I32 undef %173:sreg_32, 31, implicit-def dead $scc
+  ; CHECK-NEXT:   undef %178.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_]], undef %173:sreg_32, implicit-def $scc
+  ; CHECK-NEXT:   %178.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_]], [[S_ASHR_I32_3]], implicit-def dead $scc, implicit $scc
+  ; CHECK-NEXT:   undef %185.sub0:sreg_64 = S_ADD_U32 %50.sub0, [[S_LSHL_B32_]], implicit-def $scc
+  ; CHECK-NEXT:   %185.sub1:sreg_64 = S_ADDC_U32 undef %51:sreg_32, [[S_ASHR_I32_]], implicit-def dead $scc, implicit $scc
+  ; CHECK-NEXT:   undef %192.sub0:sreg_64 = S_ADD_U32 %50.sub0, [[S_LSHL_B32_1]], implicit-def $scc
+  ; CHECK-NEXT:   %192.sub1:sreg_64 = S_ADDC_U32 undef %51:sreg_32, [[S_ASHR_I32_1]], implicit-def dead $scc, implicit $scc
+  ; CHECK-NEXT:   undef %202.sub0:sreg_64 = S_ADD_U32 %50.sub0, undef %173:sreg_32, implicit-def $scc
+  ; CHECK-NEXT:   %202.sub1:sreg_64 = S_ADDC_U32 undef %51:sreg_32, [[S_ASHR_I32_3]], implicit-def dead $scc, implicit $scc
   ; CHECK-NEXT:   [[S_ADD_U32_1:%[0-9]+]]:sreg_32 = S_ADD_U32 %50.sub0, 224, implicit-def $scc
   ; CHECK-NEXT:   [[S_ADDC_U32_1:%[0-9]+]]:sreg_32 = S_ADDC_U32 undef %51:sreg_32, 0, implicit-def dead $scc, implicit $scc
-  ; CHECK-NEXT:   undef %210.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_1]], [[S_LSHL_B32_]], implicit-def $scc
-  ; CHECK-NEXT:   %210.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_1]], [[S_ASHR_I32_]], implicit-def dead $scc, implicit $scc
-  ; CHECK-NEXT:   undef %217.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_1]], [[S_LSHL_B32_1]], implicit-def $scc
-  ; CHECK-NEXT:   %217.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_1]], [[S_ASHR_I32_1]], implicit-def dead $scc, implicit $scc
-  ; CHECK-NEXT:   undef %224.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_1]], [[S_LSHL_B32_2]], implicit-def $scc
-  ; CHECK-NEXT:   %224.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_1]], [[S_ASHR_I32_2]], implicit-def dead $scc, implicit $scc
+  ; CHECK-NEXT:   undef %212.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_1]], [[S_LSHL_B32_]], implicit-def $scc
+  ; CHECK-NEXT:   %212.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_1]], [[S_ASHR_I32_]], implicit-def dead $scc, implicit $scc
+  ; CHECK-NEXT:   undef %219.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_1]], [[S_LSHL_B32_1]], implicit-def $scc
+  ; CHECK-NEXT:   %219.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_1]], [[S_ASHR_I32_1]], implicit-def dead $scc, implicit $scc
+  ; CHECK-NEXT:   undef %226.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_1]], [[S_LSHL_B32_2]], implicit-def $scc
+  ; CHECK-NEXT:   %226.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_1]], [[S_ASHR_I32_2]], implicit-def dead $scc, implicit $scc
   ; CHECK-NEXT:   [[S_ADD_U32_2:%[0-9]+]]:sreg_32 = S_ADD_U32 %50.sub0, 576, implicit-def $scc
   ; CHECK-NEXT:   [[S_ADDC_U32_2:%[0-9]+]]:sreg_32 = S_ADDC_U32 undef %51:sreg_32, 0, implicit-def dead $scc, implicit $scc
-  ; CHECK-NEXT:   undef %241.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_2]], [[S_LSHL_B32_]], implicit-def $scc
-  ; CHECK-NEXT:   %241.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_2]], [[S_ASHR_I32_]], implicit-def dead $scc, implicit $scc
-  ; CHECK-NEXT:   undef %253.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_2]], [[S_LSHL_B32_2]], implicit-def $scc
-  ; CHECK-NEXT:   %253.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_2]], [[S_ASHR_I32_2]], implicit-def dead $scc, implicit $scc
-  ; CHECK-NEXT:   undef %261.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_2]], undef %171:sreg_32, implicit-def $scc
-  ; CHECK-NEXT:   %261.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_2]], [[S_ASHR_I32_3]], implicit-def dead $scc, implicit $scc
-  ; CHECK-NEXT:   undef %273.sub0:sreg_64 = S_ADD_U32 [[COPY6]], [[S_LSHL_B32_]], implicit-def $scc
-  ; CHECK-NEXT:   %273.sub1:sreg_64 = S_ADDC_U32 undef %48:sreg_32, [[S_ASHR_I32_]], implicit-def dead $scc, implicit $scc
-  ; CHECK-NEXT:   undef %286.sub0:sreg_64 = S_ADD_U32 [[COPY7]], [[S_LSHL_B32_1]], implicit-def $scc
-  ; CHECK-NEXT:   %286.sub1:sreg_64 = S_ADDC_U32 undef %45:sreg_32, [[S_ASHR_I32_1]], implicit-def dead $scc, implicit $scc
-  ; CHECK-NEXT:   undef %293.sub0:sreg_64 = S_ADD_U32 [[COPY7]], [[S_LSHL_B32_2]], implicit-def $scc
-  ; CHECK-NEXT:   %293.sub1:sreg_64 = S_ADDC_U32 undef %45:sreg_32, [[S_ASHR_I32_2]], implicit-def dead $scc, implicit $scc
+  ; CHECK-NEXT:   undef %243.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_2]], [[S_LSHL_B32_]], implicit-def $scc
+  ; CHECK-NEXT:   %243.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_2]], [[S_ASHR_I32_]], implicit-def dead $scc, implicit $scc
+  ; CHECK-NEXT:   undef %255.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_2]], [[S_LSHL_B32_2]], implicit-def $scc
+  ; CHECK-NEXT:   %255.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_2]], [[S_ASHR_I32_2]], implicit-def dead $scc, implicit $scc
+  ; CHECK-NEXT:   undef %263.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_2]], undef %173:sreg_32, implicit-def $scc
+  ; CHECK-NEXT:   %263.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_2]], [[S_ASHR_I32_3]], implicit-def dead $scc, implicit $scc
+  ; CHECK-NEXT:   undef %275.sub0:sreg_64 = S_ADD_U32 [[COPY6]], [[S_LSHL_B32_]], implicit-def $scc
+  ; CHECK-NEXT:   %275.sub1:sreg_64 = S_ADDC_U32 undef %48:sreg_32, [[S_ASHR_I32_]], implicit-def dead $scc, implicit $scc
+  ; CHECK-NEXT:   undef %288.sub0:sreg_64 = S_ADD_U32 [[COPY7]], [[S_LSHL_B32_1]], implicit-def $scc
+  ; CHECK-NEXT:   %288.sub1:sreg_64 = S_ADDC_U32 undef %45:sreg_32, [[S_ASHR_I32_1]], implicit-def dead $scc, implicit $scc
+  ; CHECK-NEXT:   undef %295.sub0:sreg_64 = S_ADD_U32 [[COPY7]], [[S_LSHL_B32_2]], implicit-def $scc
+  ; CHECK-NEXT:   %295.sub1:sreg_64 = S_ADDC_U32 undef %45:sreg_32, [[S_ASHR_I32_2]], implicit-def dead $scc, implicit $scc
   ; CHECK-NEXT:   [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[S_LSHL_B32_]], 16, implicit-def dead $scc
   ; CHECK-NEXT:   [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 [[S_LSHL_B32_2]], 16, implicit-def dead $scc
-  ; CHECK-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR %302, [[S_ADD_I32_]], 0 :: (dereferenceable invariant load (s32))
-  ; CHECK-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR1:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR %302, undef %314:sreg_32, 0 :: (dereferenceable invariant load (s32))
-  ; CHECK-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR2:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR %302, [[S_ADD_I32_1]], 0 :: (dereferenceable invariant load (s32))
-  ; CHECK-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM2:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM %302, 16, 0 :: (dereferenceable invariant load (s32))
+  ; CHECK-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR %304, [[S_ADD_I32_]], 0 :: (dereferenceable invariant load (s32))
+  ; CHECK-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR1:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR %304, undef %316:sreg_32, 0 :: (dereferenceable invariant load (s32))
+  ; CHECK-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR2:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR %304, [[S_ADD_I32_1]], 0 :: (dereferenceable invariant load (s32))
+  ; CHECK-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM2:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM %304, 16, 0 :: (dereferenceable invariant load (s32))
   ; CHECK-NEXT:   [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET undef %118:sgpr_128, 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
-  ; CHECK-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR3:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR undef %369:sgpr_128, undef %370:sreg_32, 0 :: (dereferenceable invariant load (s32))
-  ; CHECK-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM3:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM undef %380:sgpr_128, 16, 0 :: (dereferenceable invariant load (s32))
-  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM3:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %156, 0, 0 :: (invariant load (s128) from %ir.92, addrspace 4)
-  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM4:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %163, 0, 0 :: (invariant load (s128) from %ir.97, addrspace 4)
-  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM5:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %176, 0, 0 :: (invariant load (s128) from %ir.104, addrspace 4)
-  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM6:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %183, 0, 0 :: (invariant load (s128) from %ir.109, addrspace 4)
-  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM7:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %190, 0, 0 :: (invariant load (s128) from %ir.114, addrspace 4)
+  ; CHECK-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM3:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM undef %134:sgpr_128, 16, 0 :: (dereferenceable invariant load (s32))
+  ; CHECK-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR3:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR undef %371:sgpr_128, undef %372:sreg_32, 0 :: (dereferenceable invariant load (s32))
+  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM3:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %158, 0, 0 :: (invariant load (s128) from %ir.92, addrspace 4)
+  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM4:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %165, 0, 0 :: (invariant load (s128) from %ir.97, addrspace 4)
+  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM5:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %178, 0, 0 :: (invariant load (s128) from %ir.104, addrspace 4)
+  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM6:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %185, 0, 0 :: (invariant load (s128) from %ir.109, addrspace 4)
+  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM7:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %192, 0, 0 :: (invariant load (s128) from %ir.114, addrspace 4)
   ; CHECK-NEXT:   [[BUFFER_LOAD_FORMAT_X_IDXEN2:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_IDXEN [[V_MOV_B32_e32_]], [[S_LOAD_DWORDX4_IMM2]], 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
-  ; CHECK-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR4:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR undef %364:sgpr_128, [[S_ADD_I32_]], 0 :: (dereferenceable invariant load (s32))
-  ; CHECK-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR5:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR undef %375:sgpr_128, [[S_ADD_I32_1]], 0 :: (dereferenceable invariant load (s32))
+  ; CHECK-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR4:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR undef %366:sgpr_128, [[S_ADD_I32_]], 0 :: (dereferenceable invariant load (s32))
+  ; CHECK-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR5:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR undef %377:sgpr_128, [[S_ADD_I32_1]], 0 :: (dereferenceable invariant load (s32))
   ; CHECK-NEXT:   [[S_ADD_I32_2:%[0-9]+]]:sreg_32 = S_ADD_I32 [[S_BUFFER_LOAD_DWORD_SGPR]], -98, implicit-def dead $scc
   ; CHECK-NEXT:   [[S_ADD_I32_3:%[0-9]+]]:sreg_32 = S_ADD_I32 [[S_BUFFER_LOAD_DWORD_SGPR1]], -114, implicit-def dead $scc
   ; CHECK-NEXT:   [[S_ADD_I32_4:%[0-9]+]]:sreg_32 = S_ADD_I32 [[S_BUFFER_LOAD_DWORD_SGPR2]], -130, implicit-def dead $scc
   ; CHECK-NEXT:   [[S_ADD_I32_5:%[0-9]+]]:sreg_32 = S_ADD_I32 [[S_BUFFER_LOAD_DWORD_IMM2]], -178, implicit-def dead $scc
-  ; CHECK-NEXT:   undef %327.sub0:sreg_64 = S_ADD_U32 [[COPY8]], [[S_LSHL_B32_]], implicit-def $scc
-  ; CHECK-NEXT:   %327.sub1:sreg_64 = S_ADDC_U32 undef %42:sreg_32, [[S_ASHR_I32_]], implicit-def dead $scc, implicit $scc
-  ; CHECK-NEXT:   undef %335.sub0:sreg_64 = S_ADD_U32 [[COPY9]], [[S_LSHL_B32_]], implicit-def $scc
-  ; CHECK-NEXT:   %335.sub1:sreg_64 = S_ADDC_U32 undef %39:sreg_32, [[S_ASHR_I32_]], implicit-def dead $scc, implicit $scc
-  ; CHECK-NEXT:   undef %343.sub0:sreg_64 = S_ADD_U32 [[COPY9]], [[S_LSHL_B32_1]], implicit-def $scc
-  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM8:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %200, 0, 0 :: (invariant load (s128) from %ir.121, addrspace 4)
-  ; CHECK-NEXT:   %343.sub1:sreg_64 = S_ADDC_U32 undef %39:sreg_32, [[S_ASHR_I32_1]], implicit-def dead $scc, implicit $scc
-  ; CHECK-NEXT:   undef %351.sub0:sreg_64 = S_ADD_U32 [[COPY9]], [[S_LSHL_B32_2]], implicit-def $scc
-  ; CHECK-NEXT:   %351.sub1:sreg_64 = S_ADDC_U32 undef %39:sreg_32, [[S_ASHR_I32_2]], implicit-def dead $scc, implicit $scc
+  ; CHECK-NEXT:   undef %329.sub0:sreg_64 = S_ADD_U32 [[COPY8]], [[S_LSHL_B32_]], implicit-def $scc
+  ; CHECK-NEXT:   %329.sub1:sreg_64 = S_ADDC_U32 undef %42:sreg_32, [[S_ASHR_I32_]], implicit-def dead $scc, implicit $scc
+  ; CHECK-NEXT:   undef %337.sub0:sreg_64 = S_ADD_U32 [[COPY9]], [[S_LSHL_B32_]], implicit-def $scc
+  ; CHECK-NEXT:   %337.sub1:sreg_64 = S_ADDC_U32 undef %39:sreg_32, [[S_ASHR_I32_]], implicit-def dead $scc, implicit $scc
+  ; CHECK-NEXT:   undef %345.sub0:sreg_64 = S_ADD_U32 [[COPY9]], [[S_LSHL_B32_1]], implicit-def $scc
+  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM8:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %202, 0, 0 :: (invariant load (s128) from %ir.121, addrspace 4)
+  ; CHECK-NEXT:   %345.sub1:sreg_64 = S_ADDC_U32 undef %39:sreg_32, [[S_ASHR_I32_1]], implicit-def dead $scc, implicit $scc
+  ; CHECK-NEXT:   undef %353.sub0:sreg_64 = S_ADD_U32 [[COPY9]], [[S_LSHL_B32_2]], implicit-def $scc
+  ; CHECK-NEXT:   %353.sub1:sreg_64 = S_ADDC_U32 undef %39:sreg_32, [[S_ASHR_I32_2]], implicit-def dead $scc, implicit $scc
   ; CHECK-NEXT:   [[S_LSHL_B32_3:%[0-9]+]]:sreg_32 = S_LSHL_B32 [[COPY10]], 4, implicit-def dead $scc
   ; CHECK-NEXT:   [[BUFFER_LOAD_FORMAT_X_IDXEN3:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_IDXEN [[V_MOV_B32_e32_]], [[S_LOAD_DWORDX4_IMM3]], 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
   ; CHECK-NEXT:   [[S_ADD_I32_6:%[0-9]+]]:sreg_32 = S_ADD_I32 [[S_LSHL_B32_3]], 16, implicit-def dead $scc
   ; CHECK-NEXT:   [[S_BUFFER_LOAD_DWORD_SGPR6:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR undef %396:sgpr_128, [[S_ADD_I32_6]], 0 :: (dereferenceable invariant load (s32))
   ; CHECK-NEXT:   [[BUFFER_LOAD_FORMAT_X_IDXEN4:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_IDXEN [[V_MOV_B32_e32_]], [[S_LOAD_DWORDX4_IMM4]], 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
   ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM9:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %50, 224, 0 :: (invariant load (s128) from %ir.126, addrspace 4)
-  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM10:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %210, 0, 0 :: (invariant load (s128) from %ir.127, addrspace 4)
+  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM10:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %212, 0, 0 :: (invariant load (s128) from %ir.127, addrspace 4)
   ; CHECK-NEXT:   [[BUFFER_LOAD_FORMAT_X_IDXEN5:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_IDXEN [[V_MOV_B32_e32_]], [[S_LOAD_DWORDX4_IMM5]], 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
-  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM11:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %217, 0, 0 :: (invariant load (s128) from %ir.132, addrspace 4)
-  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM12:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %224, 0, 0 :: (invariant load (s128) from %ir.137, addrspace 4)
+  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM11:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %219, 0, 0 :: (invariant load (s128) from %ir.132, addrspace 4)
+  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM12:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %226, 0, 0 :: (invariant load (s128) from %ir.137, addrspace 4)
   ; CHECK-NEXT:   [[BUFFER_LOAD_FORMAT_X_IDXEN6:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_IDXEN [[V_MOV_B32_e32_]], [[S_LOAD_DWORDX4_IMM6]], 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
   ; CHECK-NEXT:   [[BUFFER_LOAD_FORMAT_X_IDXEN7:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_IDXEN [[V_MOV_B32_e32_]], [[S_LOAD_DWORDX4_IMM7]], 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
   ; CHECK-NEXT:   [[BUFFER_LOAD_FORMAT_X_IDXEN8:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_IDXEN [[V_MOV_B32_e32_]], [[S_LOAD_DWORDX4_IMM8]], 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
   ; CHECK-NEXT:   [[S_ADD_I32_7:%[0-9]+]]:sreg_32 = S_ADD_I32 [[S_BUFFER_LOAD_DWORD_SGPR4]], -217, implicit-def dead $scc
   ; CHECK-NEXT:   [[S_ADD_I32_8:%[0-9]+]]:sreg_32 = S_ADD_I32 [[S_BUFFER_LOAD_DWORD_SGPR3]], -233, implicit-def dead $scc
   ; CHECK-NEXT:   [[S_ADD_I32_9:%[0-9]+]]:sreg_32 = S_ADD_I32 [[S_BUFFER_LOAD_DWORD_SGPR5]], -249, implicit-def dead $scc
   ; CHECK-NEXT:   [[S_ADD_I32_10:%[0-9]+]]:sreg_32 = S_ADD_I32 [[S_BUFFER_LOAD_DWORD_IMM3]], -297, implicit-def dead $scc
   ; CHECK-NEXT:   [[S_ADD_I32_11:%[0-9]+]]:sreg_32 = S_ADD_I32 [[S_BUFFER_LOAD_DWORD_SGPR3]], -313, implicit-def dead $scc
   ; CHECK-NEXT:   [[S_ADD_I32_12:%[0-9]+]]:sreg_32 = S_ADD_I32 [[S_BUFFER_LOAD_DWORD_SGPR3]], -329, implicit-def dead $scc
   ; CHECK-NEXT:   [[S_ADD_I32_13:%[0-9]+]]:sreg_32 = S_ADD_I32 [[S_BUFFER_LOAD_DWORD_SGPR3]], -345, implicit-def dead $scc
   ; CHECK-NEXT:   [[S_ADD_I32_14:%[0-9]+]]:sreg_32 = S_ADD_I32 [[S_BUFFER_LOAD_DWORD_SGPR6]], -441, implicit-def dead $scc
   ; CHECK-NEXT:   [[S_ADD_U32_3:%[0-9]+]]:sreg_32 = S_ADD_U32 [[COPY1]], 160, implicit-def $scc
   ; CHECK-NEXT:   [[S_ADDC_U32_3:%[0-9]+]]:sreg_32 = S_ADDC_U32 undef %36:sreg_32, 0, implicit-def dead $scc, implicit $scc
   ; CHECK-NEXT:   undef %411.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_3]], [[S_LSHL_B32_2]], implicit-def $scc
   ; CHECK-NEXT:   %411.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_3]], [[S_ASHR_I32_2]], implicit-def dead $scc, implicit $scc
   ; CHECK-NEXT:   [[S_LSHL_B32_4:%[0-9]+]]:sreg_32 = S_LSHL_B32 [[COPY11]], 4, implicit-def dead $scc
   ; CHECK-NEXT:   [[BUFFER_LOAD_FORMAT_X_IDXEN9:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_IDXEN [[V_MOV_B32_e32_]], [[S_LOAD_DWORDX4_IMM10]], 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
   ; CHECK-NEXT:   [[S_ASHR_I32_4:%[0-9]+]]:sreg_32_xm0 = S_ASHR_I32 [[S_LSHL_B32_4]], 31, implicit-def dead $scc
   ; CHECK-NEXT:   undef %425.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_3]], [[S_LSHL_B32_4]], implicit-def $scc
   ; CHECK-NEXT:   %425.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_3]], [[S_ASHR_I32_4]], implicit-def dead $scc, implicit $scc
   ; CHECK-NEXT:   [[S_ADD_U32_4:%[0-9]+]]:sreg_32 = S_ADD_U32 %56.sub0, 168, implicit-def $scc
   ; CHECK-NEXT:   [[S_ADDC_U32_4:%[0-9]+]]:sreg_32 = S_ADDC_U32 undef %57:sreg_32, 0, implicit-def dead $scc, implicit $scc
-  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM13:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %241, 0, 0 :: (invariant load (s128) from %ir.147, addrspace 4)
+  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM13:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %243, 0, 0 :: (invariant load (s128) from %ir.147, addrspace 4)
   ; CHECK-NEXT:   [[S_LSHL_B32_5:%[0-9]+]]:sreg_32 = S_LSHL_B32 [[COPY4]], 3, implicit-def dead $scc
   ; CHECK-NEXT:   [[BUFFER_LOAD_FORMAT_X_IDXEN10:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_IDXEN [[V_MOV_B32_e32_]], [[S_LOAD_DWORDX4_IMM11]], 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
   ; CHECK-NEXT:   [[S_ASHR_I32_5:%[0-9]+]]:sreg_32_xm0 = S_ASHR_I32 [[S_LSHL_B32_5]], 31, implicit-def dead $scc
   ; CHECK-NEXT:   undef %441.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_4]], [[S_LSHL_B32_5]], implicit-def $scc
   ; CHECK-NEXT:   %441.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_4]], [[S_ASHR_I32_5]], implicit-def dead $scc, implicit $scc
   ; CHECK-NEXT:   [[S_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_LOAD_DWORD_IMM %441, 0, 0 :: (invariant load (s32) from %ir.269, align 8, addrspace 4)
-  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM14:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %253, 0, 0 :: (invariant load (s128) from %ir.154, addrspace 4)
+  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM14:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %255, 0, 0 :: (invariant load (s128) from %ir.154, addrspace 4)
   ; CHECK-NEXT:   [[BUFFER_LOAD_FORMAT_X_IDXEN11:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_IDXEN [[V_MOV_B32_e32_]], [[S_LOAD_DWORDX4_IMM12]], 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
-  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM15:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %261, 0, 0 :: (invariant load (s128) from %ir.159, addrspace 4)
+  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM15:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %263, 0, 0 :: (invariant load (s128) from %ir.159, addrspace 4)
   ; CHECK-NEXT:   [[BUFFER_LOAD_FORMAT_X_IDXEN12:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_IDXEN [[V_MOV_B32_e32_]], [[S_LOAD_DWORDX4_IMM9]], 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
   ; CHECK-NEXT:   [[BUFFER_LOAD_FORMAT_X_IDXEN13:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_IDXEN [[V_MOV_B32_e32_]], [[S_LOAD_DWORDX4_IMM13]], 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
   ; CHECK-NEXT:   %71.sub3:sgpr_128 = S_MOV_B32 553734060
   ; CHECK-NEXT:   %71.sub2:sgpr_128 = S_MOV_B32 -1
   ; CHECK-NEXT:   [[COPY13:%[0-9]+]]:sgpr_128 = COPY %71
-  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM16:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %273, 0, 0 :: (invariant load (s128) from %ir.167, addrspace 4)
-  ; CHECK-NEXT:   [[COPY13]].sub1:sgpr_128 = COPY %302.sub1
+  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM16:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %275, 0, 0 :: (invariant load (s128) from %ir.167, addrspace 4)
+  ; CHECK-NEXT:   [[COPY13]].sub1:sgpr_128 = COPY %304.sub1
   ; CHECK-NEXT:   [[COPY13]].sub0:sgpr_128 = COPY [[S_LOAD_DWORD_IMM]]
   ; CHECK-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM4:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[COPY13]], 0, 0 :: (dereferenceable invariant load (s32))
   ; CHECK-NEXT:   [[BUFFER_LOAD_FORMAT_X_IDXEN14:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_IDXEN [[V_MOV_B32_e32_]], [[S_LOAD_DWORDX4_IMM14]], 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
   ; CHECK-NEXT:   [[BUFFER_LOAD_FORMAT_X_IDXEN15:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_IDXEN [[V_MOV_B32_e32_]], [[S_LOAD_DWORDX4_IMM15]], 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
-  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM17:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %286, 0, 0 :: (invariant load (s128) from %ir.175, addrspace 4)
-  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM18:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %293, 0, 0 :: (invariant load (s128) from %ir.180, addrspace 4)
+  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM17:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %288, 0, 0 :: (invariant load (s128) from %ir.175, addrspace 4)
+  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM18:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %295, 0, 0 :: (invariant load (s128) from %ir.180, addrspace 4)
   ; CHECK-NEXT:   [[BUFFER_LOAD_FORMAT_X_IDXEN16:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_IDXEN [[V_MOV_B32_e32_]], [[S_LOAD_DWORDX4_IMM16]], 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
   ; CHECK-NEXT:   [[S_LSHL_B32_6:%[0-9]+]]:sreg_32 = S_LSHL_B32 [[COPY3]], 3, implicit-def dead $scc
   ; CHECK-NEXT:   [[BUFFER_LOAD_DWORD_OFFSET1:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[S_LOAD_DWORDX4_IMM1]], 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
   ; CHECK-NEXT:   [[S_ASHR_I32_6:%[0-9]+]]:sreg_32_xm0 = S_ASHR_I32 [[S_LSHL_B32_6]], 31, implicit-def dead $scc
   ; CHECK-NEXT:   [[S_ADD_I32_15:%[0-9]+]]:sreg_32 = S_ADD_I32 [[S_BUFFER_LOAD_DWORD_IMM4]], -467, implicit-def dead $scc
   ; CHECK-NEXT:   undef %453.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_4]], [[S_LSHL_B32_6]], implicit-def $scc
   ; CHECK-NEXT:   %453.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_4]], [[S_ASHR_I32_6]], implicit-def dead $scc, implicit $scc
   ; CHECK-NEXT:   [[S_LOAD_DWORDX2_IMM:%[0-9]+]]:sreg_64_xexec = S_LOAD_DWORDX2_IMM %453, 0, 0 :: (invariant load (s64) from %ir.277, addrspace 4)
   ; CHECK-NEXT:   [[BUFFER_LOAD_DWORD_OFFSET2:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[S_LOAD_DWORDX4_IMM17]], 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
   ; CHECK-NEXT:   [[BUFFER_LOAD_DWORD_OFFSET3:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[S_LOAD_DWORDX4_IMM18]], 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
-  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM19:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %327, 0, 0 :: (invariant load (s128) from %ir.202, addrspace 4)
-  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM20:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %335, 0, 0 :: (invariant load (s128) from %ir.208, addrspace 4)
+  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM19:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %329, 0, 0 :: (invariant load (s128) from %ir.202, addrspace 4)
+  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM20:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %337, 0, 0 :: (invariant load (s128) from %ir.208, addrspace 4)
   ; CHECK-NEXT:   [[COPY14:%[0-9]+]]:sgpr_128 = COPY %71
-  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM21:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %343, 0, 0 :: (invariant load (s128) from %ir.213, addrspace 4)
+  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM21:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %345, 0, 0 :: (invariant load (s128) from %ir.213, addrspace 4)
   ; CHECK-NEXT:   [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[S_LOAD_DWORDX2_IMM]].sub1, 65535, implicit-def dead $scc
   ; CHECK-NEXT:   [[COPY14]].sub0:sgpr_128 = COPY [[S_LOAD_DWORDX2_IMM]].sub0
   ; CHECK-NEXT:   [[COPY14]].sub1:sgpr_128 = COPY [[S_AND_B32_]]
   ; CHECK-NEXT:   [[S_BUFFER_LOAD_DWORD_IMM5:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[COPY14]], 0, 0 :: (dereferenceable invariant load (s32))
-  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM22:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %351, 0, 0 :: (invariant load (s128) from %ir.218, addrspace 4)
+  ; CHECK-NEXT:   [[S_LOAD_DWORDX4_IMM22:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM %353, 0, 0 :: (invariant load (s128) from %ir.218, addrspace 4)
   ; CHECK-NEXT:   [[BUFFER_LOAD_FORMAT_X_IDXEN17:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_IDXEN [[V_MOV_B32_e32_]], [[S_LOAD_DWORDX4_IMM19]], 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
   ; CHECK-NEXT:   [[BUFFER_LOAD_FORMAT_X_IDXEN18:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_IDXEN [[V_MOV_B32_e32_]], [[S_LOAD_DWORDX4_IMM20]], 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
   ; CHECK-NEXT:   [[S_LSHL_B32_7:%[0-9]+]]:sreg_32 = S_LSHL_B32 [[COPY2]], 3, implicit-def dead $scc
   ; CHECK-NEXT:   [[BUFFER_LOAD_FORMAT_X_IDXEN19:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_IDXEN [[V_MOV_B32_e32_]], [[S_LOAD_DWORDX4_IMM21]], 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 7)
   ; CHECK-NEXT:   [[S_ASHR_I32_7:%[0-9]+]]:sreg_32_xm0 = S_ASHR_I32 [[S_LSHL_B32_7]], 31, implicit-def dead $scc
   ; CHECK-NEXT:   [[S_ADD_I32_16:%[0-9]+]]:sreg_32 = S_ADD_I32 [[S_BUFFER_LOAD_DWORD_IMM5]], -468, implicit-def dead $scc
   ; CHECK-NEXT:   undef %468.sub0:sreg_64 = S_ADD_U32 [[S_ADD_U32_4]], [[S_LSHL_B32_7]], implicit-def $scc
   ; CHECK-NEXT:   %468.sub1:sreg_64 = S_ADDC_U32 [[S_ADDC_U32_4]], [[S_ASHR_I32_7]], implicit-def dead $scc, implicit $scc